- Fix S3 acceess by dropping aiobotocore patching and using public API instead. (#3155)
No significant changes.
-
Added cluster-bound apps configuration into the SDK. (#3142)
-
Prefer
.apoloconfiguration folder and.apoloignorefiles and folders over old.neurobased names; old names are still supported. (#3145) -
Balance is no longer stored on a cluster level, and was moved to an organization level, e.g., to an org itself, and to an org users, instead of a cluster / cluster users.
New commands:
apolo admin set-org-defaults- allows to set an organization default values, such as a default user credits
Existing commands changes:
apolo admin add-cluster-usercmd is no longer accepting acreditsargument.apolo admin set-user-creditscmd is now expecting an org name instead of a cluster name.apolo admin add-user-creditscmd is now expecting an org name instead of a cluster name.apolo admin set-org-cluster-creditswas removed in a favor of anapolo admin set-org-credits.apolo admin add-org-cluster-creditswas removed in a favor of anapolo admin add-org-credits. (#3137)
No significant changes.
No significant changes.
No significant changes.
No significant changes.
No significant changes.
No significant changes.
No significant changes.
No significant changes.
No significant changes.
No significant changes.
No significant changes.
No significant changes.
No significant changes.
No significant changes.
No significant changes.
- Publish neuro-extras for backward compatibility too.
No significant changes.
Renamed packages from Neuro to Apolo
No significant changes.
- Rename neuro-extras to apolo-extras. (#605)
No significant changes.
- Fix image cache path -- properly support orgs (#604)
- Fix image build in organization when user has no direct access to cluster (#603)
No significant changes.
- Support AMD, Intel GPUs (#3007)
No significant changes.
- Updated Kaniko to 1.20.0, allowed to provide extra arguments for Kaniko executor. (#601)
-
${{ project.project_name }}now also configuresvolume's remote path andimagereference if the project name was not set.If you do not have
project_nameset inproject.yaml, the volume paths are assumed within your current project configured in CLI. However, if you setproject_name, this project will be assumed while managing jobs, bakes, volumes, building images etc. within this flow. (#1081) -
Added support of extra Kaniko arguments while building an image:
images: image_a: ref: image:imagea context: dir dockerfile: dir/Dockerfile extra_kaniko_args: >- --reproducible --cache-ttl=1h --single-snapshotMore details on available arguments could be found in official Kaniko documentation. (#1110)
- Do not perform terminal settings reset in PyCharm while it causes screen cleanup. (#3005)
No significant changes.
No significant changes.
- Use
-p/--projectwhile building image to set the project for entire build process. (#600)
-
${{ project.project_name }}now also configuresvolume's remote path andimagereference if the project name was not set.If you do not have
project_nameset inproject.yaml, the volume paths are assumed within your current project configured in CLI. However, if you setproject_name, this project will be assumed while managing jobs, bakes, volumes, building images etc. within this flow. (#1081) -
Add retries for read-only operations on Server Unavailable error. (#1085)
- Fix read retries from neuro APIs. Fix configuration of executor lifespan in bake config file, example:
kind: batch life_span: 30d tasks: ... ``` ([#1075](https://github.com/neuro-inc/neuro-flow/issues/1075))
- Expose jobs and tasks restart policy configuration (#1072)
- Fix runnning Kaniko on some envs (#599)
- Commands
neuro run,neuro logs,neuro attachandneuro execin non-quiet mode now prints details for cancelled and failed jobs. Also improved other indications of the job status. (#2800) - Commands
neuro runandneuro attachnow re-attach to restarted jobs. Previously they waited until the job finished. (#2802) - Added support of cluster energy schedules.
neuro config show --energywill display awailable energy schedule periods.neuro run --schedule-name <energy-schedule-name>will run the job within the specified . Note - the selected preset should have an enabled scheduler.neuro status <job-id>will include the energy schedule name if one was used for running the job. (#2903) - Added Python 3.11 support. (#2908)
- Add
neuro config switch-projectcommand. (#2940) - Support projects in storage commands, make project directory default instead of user home directory. (#2949)
- Support projects in image commands. (#2950)
- Support projects in
neuro secretcli. (#2952) - Support projects in
neuro blobcli commands andneuro-sdk. (#2954) - Add project support in
neuro jobcommands. (#2955) - Support projects in disk CLI commands and SDK methods. (#2960)
- Support projects in service account CLI commands and SDK methods. (#2965)
- Unhide remove-project command. (#2976)
- Align all
lscommand options. (#2995)
- Fix human-readable datetime formatter to correctly report jobs lifespan ending in future (#2933)
- Update projects after fetching config from server. (#2992)
- Added project support. ([#595])(neuro-inc/apolo-extras#595))
-
Fixed data cp to s3 destination not working. (#301)
-
Fixed a bug with data cp to s3 with compression causing
$HOME/directory wipe. (#500)
- Dropped Python 3.7, added 3.10, 3.11 support. (#980)
- Host neuro-flow flows (former projects) within user projects. (#1002)
- Added
hash_files_relativefunction to expression, it works same ashash_filesbut requires additional leading parameters that defines directory to glob over. It can be used to glob over action files:${{ hash_files_relative(flow.action_path, "**/pattern/here/**/*.py", "other/**/pattern") ``` ([#904](https://github.com/neuro-inc/neuro-flow/issues/904))
- Added support of
flow.action_pathfor images sections of action. (#902)
- Add update-node-pool command. Only idle_size update is available. (#2740)
- Integrate neuro-config-client package. (#2791)
- Added project management commands to
neuro admin. (#2798)
- Fix
neuro admin update-resource-preset: previous values are now used for boolean parameters if options--scheduler/--no-scheduleror--preemptible-node/--no-preemptible-nodeare not explicitly specified. (#2761)
- Implement
flow.action_pathproperty in the flow context. It is available from action and points to the folder whereaction.ymlfile is located. (#896) - Replace HTTP fetch with git clone for remote actions. (#897)
- Use a separate src folder to don install tests when installing neuro-flow. (#891)
- Implement
neuro-flow initcommand for easy flow creation (#859)
- Update size formatting: use decimal prefixes (#2768)
- Fix formatting URIs in short form for users with organization. (#2749)
- Disks, buckets and secrets are now created with the current organization instead of no organization if
--orgis not explicitly specified. (#2756)
- Pin
neuro-admin-clientto>=22.6.4, make SDK future-compatible with nextneuro-admin-clientreleases. (#2757)
- Allow user without clusters call neuro-sdk Jobs.status method if they have permissions. (#2733)
- Re-load clusters config in
neuro config show. (#2728)
-
Support organization storage in disk usage command when user switches to organization. (#2585)
-
Added
neuro admin set-org-cluster-defaultscommand to allow org cluster manager to modify org cluster defaults. (#2602) -
Support custom storage path in disk usage command. (#2622)
-
neuro blob statbucketnow displays some additional information for imported buckets.- For AWS: "External name", "External endpoint", "External region name".
- For Azure: "External name", "External endpoint".
- For GCP: "External name". (#2706)
- Fix resolving job URIs containing organization name and/or service account name. (#2572)
- Provide click 8.1.3 support with keeping 8.0+ compatibility. (#2689)
- Add support of shared server side projects. (#840)
- Display times with timezones in log messages. (#847)
- Fixed error when trying to build image in batch mode (#818)
- Fix relative dockerfile path for Kaniko on Windows (#463)
-
Added Python 3.10 support. (#417)
-
Add
local-buildsubcommand toneuro-extras imagethat allows building images via local Docker daemon. (#448) -
Help user to select the resource preset for image build (#450)
- Removed Python 3.6 support. (#417)
No significant changes.
- Fixed problem with click 8.1.0
- Support organization storage size parameter. (#2560)
- Added client-side validation of docker image name and tag. (#2525)
- Added support of column type
org_nameto ps-format. (#2533) - Added "ORG" column to get-cluster-users output. (#2541)
- Added info about org quota to
neuro config show. (#2545)
- Cluster/org name validness now checked after config re-fetch in
neuro config switch-clusterandneuro config switch-orgcommands. (#2543)
-
Add command
neuro acl list-rolesto list roles created by user. (#2496) -
Added support to set/update cluster level default credits and quota.
Use options
--default-creditsand--default-jobsof commandsneuro admin add-cluster,neuro admin update-cluster,neuro admin add-org-clusterandneuro admin update-org-clusterto set and update cluster defaults. This values will when new user with role "user" is added to cluster (either byneuro admin add-cluster-useror if user registers himself using web interface). The default for managers and admins is unlimited quota and credits as the can edit their quota.You can override default value by using
--creditsand--jobsoptions ofneuro admin add-org-clustercommand. (#2520)
- Fixed memory leak in
neuro blob cp. (#2523)
-
Support
${{ matrix.ORDINAL }}as unique 0-based index for selected rows.If a batch flow has a matrix, all matrix rows are enumerated. The ordinal number of each row is available as
${{ matrix.ORDINAL }}system value. (#693) -
Add support of expressions in matrix definition.
You can now use expression to define list of matrix products. In examples here and below, both
old_way_keyandnew_way_keyproduce same list of values:matrix: old_way_key: ["1", "2", "3"] new_way_key: ${{ ["1", "2", "3"] }}This can be helpful when used together with new
range()function:matrix: old_way_key: [0, 1, 2, 3] new_way_key: ${{ range(4) }}The
range()function supports same parameters as python'srange(), but it returns list. For example:${{ range(1, 4) }}generates[1,2,3], and${{ range(4, 1, -1) }}generates[4,3,2].As sometimes plain numbers is not best options for matrix products, you can use list comprehension to perform some mapping:
matrix: old_way_key: ["k1", "k2", "k3"] new_way_key: ${{ [fmt("k{}", str(it)) for it in range(1, 4)] }}You can also filter some values in comprehension same way as in python:
matrix: old_way_key: [0, 4, 16] new_way_key: ${{ [it * it for it in range(1, 5) if it % 2 == 0] }} ``` ([#741](https://github.com/neuro-inc/neuro-flow/issues/741))
- Fixed (disabled) uploads to storage in dry-run mode (#732)
- Added support
--ownerargument in blob bucket level commands to allow referring to another users bucket by name. (#2494) - Add --force/-f flag to neuro blob rmbucket to force remove non-empty bucket. (#2495)
- Support
neuro -q lsandneuro -q blob lsfor quiet output enforcing. (#2506)
- Replace
--httpoption with--http-port, keep--httpoption as a hidden deprecated alternative, scheduled for removal later. (#2501) - Remove deprecated
neuro project initcommand. (#2502) - Remove
client.job.tags()method from SDK andneuro job tagscommand from CLI. (#2503) - Drop deprecated API from SDK (#2505)
- Sort CLI commands, groups and topics alphabetically. (#2488)
- Add support for Open-Sourced user-less services deployment. (#2492)
- Replace "Spend credits" with "Credits spent" in formatter. (#2435)
- Fix importing of asynccontextmanager from wrong place
-
Merge
config show-quotaintoconfig showCLI command. (#2436) -
Added admin commands to manager orgs and org-clusters:
Managing orgs:
neuro admin get-orgsPrint the list of available orgs user has access to.neuro admin add-org <org_name>Create a new org.neuro admin remove-org <org_name>Drop a org. Removes all memberships, very dangerous operation.
Managing org members:
neuro admin get-org-users <org_name>List all members of orgs.neuro admin add-org-user <org_name> <username>Add user to the org.neuro admin remove-org-user <org_name> <username>Remove user from the org.
Managing access of orgs to clusters:
neuro admin get-org-clusters <cluster_name>Print the list of all orgs in the clusterneuro admin add-org-cluster <cluster_name> <org_name>Add org access to specified clusterneuro admin get-org-cluster-quota <cluster_name> <org_name>Get info about org quota in given clusterneuro admin set-org-cluster-quota [options] <cluster_name> <org_name>Set org cluster quota to given valuesneuro admin set-org-cluster-credits [options] <cluster_name> <org_name>Set org cluster credits to given valueneuro admin add-org-cluster-credits [options] <cluster_name> <org_name>Add given values to org cluster balance
Manging access of org members to clusters:
neuro admin get-cluster-users --org <org_name> <cluster_name>List all members of orgs added to clusterneuro admin add-cluster-user --org <org_name> <cluster_name> <username>Add org member to clusterneuro admin remove-cluster-user --org <org_name> <cluster_name> <username>Remove org member user from the cluster.neuro admin get-user-quota --org <org_name> <cluster_name> <username>Get info about org member quota in given clusterneuro admin set-user-quota [options] --org <org_name> <cluster_name> <username>Set org member quota in cluster to given valuesneuro admin set-user-credits [options] --org <org_name> <cluster_name> <username>Set org member credits in cluster to given valueneuro admin add-user-credits [options] --org <org_name> <cluster_name> <username>Add given values to org member balance in cluster (#2449)
-
Option
-j/--jobsinneuro admin set-user-quotais now required. Passunlimitedfor setting no limit. Negative values are now rejected. (#2453) -
Option
-c/--creditsinneuro admin set-user-creditsandneuro admin add-user-creditsis now required. Pass "unlimited" inneuro admin set-user-creditsfor setting no limit. Values "infinity", "nan" etc are now rejected. (#2454) -
Added support of organizations.
Current organization is displayed in
neuro config show. It can be changed usingneuro config switch-org. To least organizations you have access to in each cluster, useneuro config get-clusters.Also you can run job on behalf of organization. By default,
neuro runwill use current organization, but you can override it usingneuro run --org <some_org>. (#2465)
- Fix command description for
neuro admin show-cluster-options. (#2434)
- Fix
neuro rm storage:...command
- Add Python 3.10 support. (#2421)
- Drop Python 3.6 support. (#2421)
- Technical release, compatible with the latest SDK/CLI.
- Configure version checker settings by plugins. (#2405)
- CLI startup time is 2 times shorter now. (#2417)
- Deprecate
neuro project initcommand, usecookiecutter gh:neuro-inc/cookiecutter-neuro-projectinstead. (#2418)
-
Allow to build and push docker images to the remote image registries.
The requirement for allowing Kaniko (the underlying tool) to push such images - it should be authenticated.
The registry authention data structure should be in the following form:
{"auths": {"<registry URI>": {"auth": "<base64 encoded '<username>:<password>' string"}}}(tested with thehttps://index.docker.io/v1/- public DockerHub registry)neuro-extras config build-registy-authcommand might become handy in this case.To attach the target registy AUTH data into the builder job, one might save it as the platform secret and mount into the builder job. Mounting of the secret could be done either as the ENV variable with the name preffixed by
NE_REGISTRY_AUTH, or as the secret file. In the later case, the ENV variable should also be added with the mentioned above preffix and pointing to the corresponding file. (#328)
- Allow
bashandpythoncode blocks in local actions. (#667) - Add
-s/--suffixoption usage hint for cases when the live job launched without it. (#679)
- Handle cached image status color in
neuro-flow inspect <bake>. (#657) - Fixed parsing of bash and python in mixins (#674)
- Fix parsing of 'bash' and 'python' usage in project config. (#678)
- Fixed logging of filename as "." for github actions. (#681)
- Add support of symlinks in
neuro ls. (#2400)
- Fix sharing buckets using
neuro acl grant. (#2414)
- Report job price in
neuro job status, add support of organisation names in jobs. (#2404) - Raise dedicated
NotSupportedErrorfor unsupported REST API calls (#2407)
-
Added new context called git. It has following attributes:
- ${{ git.sha }} -- SHA of current commit
- ${{ git.branch }} - name of current branch
- ${{ git.tags }} - list of tags that point to current commit
This context is available everywhere both in live and batch mode, but it can only be used if project's workspace is inside some git repository. (#603)
-
Added ternary operator to expressions:
${{ left if condition else right }}will evaluate toleftifconditionis True and torightotherwise. (#605) -
Allowed ']]' and '}}' in yaml literal strings. (#607)
-
Added column with batch name to output of
neuro-flow bakes. (#618) -
Enabled detailed automatic logging to file. For locally executed commands, logs will go to
~/.neuro/logsdirectory. Remote executor detailed logs will go tostorage:.flow/logs/<bake_id>/. (#622) -
Added support of non-string types for action inputs. Now, one can specify action input in following way:
inputs: arg1: descr: Input with implicit string type arg2: descr: Input with explicit string type type: str arg2: descr: Input with explicit int type and corresponding default type: int default: 42Supported types are
int,float,bool,str. On action calls side, it's now possible to use corresponding yaml types as arguments. (#626) -
Added possibility to specify job/bake params via the shortcut
-pfor--param. (#629)
-
Added support of open stack as bucket provider. (#2371)
-
Added
blob du <bucket_id_or_name>command to change bucket's storage usage. (#2372) -
Added support of bucket name in uris for bucket that do not belong to current user. (#2385)
-
Added first name, lastname, registered columns to
neuro admin get-cluster-usersoutput. (#2387) -
Added "Spent credits" to add "neuro admin get-cluster-users". Splitted 'quota' into 'balance' and 'quota':
neuro admin set-user-quotachanges onlymax_running_jobsquota.neuro admin set-user-creditschanges onlycreditsbalance of user.neuro admin add-user-creditsupdatescreditsbalance of user by delta. (#2391)
- Fix fetching zero-size blob from S3 compatible bucket providers. (#2394)
- Added support of blob storage for azure clusters. (#2284)
- Added support of blob storage for gcp clusters. (#2293)
- Added
--read-onlyflag toneuro blob mkcredentialsto allow creation of read-only credentials. (#2295) - Added
neuro blob importbucketcommand to import external bucket. External buckets support the same operations (neuro blob ls/cp/rm/glob), but it is impossible to generate persistent credentials for them using "neuro blob mkcredentials". (#2297) - Added
neuro blob sign-url <blob-uri>command for generation of urls that grant temporal access to blob. (#2299) - Added command
neuro blob set-bucket-publicity <bucket> <"public"|"private">to make bucket accessible for unauthorized users. (#2306) - Undeprecate
neuro job savecommand, we decided to support this functionality in future. (#2336)
- Revert back (#2274);
pip install neuro-cli[awscli,boto3]did not work well.
-
Implement URI autocompletion for "blob:" scheme. (#2273)
-
Provide
pip install neuro-cli[awscli,boto3]extra dependencies for installing the compatible AWS CLI version.-
neuro-cli[awscli]installsawsclipackage. -
neuro-cli[boto3]installsboto3package. -
neuro-cli[awscli,boto3]installs both. (#2274)
-
- Add support of minio provider as a bucket blob for onprem clusters (#2271)
- Relax aiobotocore requirement; it allows working with the latest aiobotocore without publishing new Neuro SDK. (#2270)
-
Add option
--sincein commandneuro logs. Add parameter since in jobs.monitor(). (#1964) -
Add option
--timestampsin commandneuro logs. Add parameter timestamps in jobs.monitor(). (#2252) -
Implement public URI helpers. (#2253)
-
Reworked blob storage support:
Bucket management commands:
- use
neuro blob mkbucketto create new bucket - use
neuro blob lsbucketto list your buckets - use
neuro blob statbucket <bucket_id>to retrieve info about single bucket - use
neuro blob rmbucket <bucket_id>to delete bucket. Note that you can only delete empty buckets.
Users can name buckets objects. The name should be unique between users buckets and can be used instead of bucket id.
Bucket contents management commands:
- use
neuro blob ls blob:<bucket_id>/<path_in_bucket>to list bucket contents. Option-rdisables file system emulation and displays all keys that start with <path_in_bucket>. - use
neuro blob glob blob:<bucket_id>/<glob_pattern>to glob search objects in buckets. For example,blob:my_bucket/**/*.txtpattern will match all.txtfiles inmy_bucketbucket. - use
neuro blob cp <src_uri> <dst_uri>to copy data from/to bucket. - use
neuro blob rm blob:<bucket_id>/<path_in_bucket>to delete elements from bucket. (#2258)
- use
-
Implement URI autocompletion for bash shell. (#2259)
-
Added way to create persistent credentials to buckets to use in outside of the platform:
- use
neuro blob mkcredentials <bucket1_id> <bucket2_id> ...to create new credentials for specified buckets - use
neuro blob lscredentialsto list your buckets credentials - use
neuro blob statcredentials <credentials_id>to retrieve info about single bucket credentials - use
neuro blob rmbucket <credentials_id>to delete bucket credentials.
Users can name buckets credentials. The name should be unique between users buckets credentials and can be used instead of bucket credentials id. (#2264)
- use
- Deprecate
neuro run IMAGE CMD, useneuro run IMAGE -- CMDinstead. The same forneuro execcommand. (#2260) - Remove deprecated options older than 6 months. (#2262)
-
Added possibility to disable PyPi version check via environment variable 'NEURO_CLI_DISABLE_PYPI_VERSION_CHECK'. (#2237)
-
Add parameter separator in jobs.monitor(). (#2239)
-
Support URI autocompletion in storage commands for ZSH shell. (#2248)
neuro savecommand is deprecated and will be removed in future CLI release. (#2249)
- Fix config permission bits for --pass-config mode (#2233)
- Drop legacy code that puts configuration files for pass-config on storage. (#2233)
- Support VCD cluster config in generate-cluster-config command. (#2193)
- Enabled writing of full logs for each CLI run to a file. Logs are generated under
~/.neuro/logsfolder and automatically removed after 3 days. (#2211) - Added
--clusteroption toneuro runcommand andneuro disk,neuro secretcommand groups. This option allows to perform actions for a specific cluster instead of current one. (#2225)
_Admin.set_user_quotamethod is fixed for cases, when the amount of credits equals to zero. (#2229)- Recover cursor after an abnormal exit (Ctrl-C). (#2231)
-
All asynchronous iterators returned by API support now an asynchronous manager protocol. It is strongly preferable to use "asyn with" before iterating them. For example::
async with client.jobs.list() as jobs: async for job in jobs: print(job.id) ([#2192](https://github.com/neuro-inc/platform-client-python/issues/2192)) -
Added
neuro storage dfcommand that allows to retrieve cluster's storage disk usage. (#2201)
- Removed additional '\n' from output of commands that print tokens. (#2200)
- Introduce NDJSONError exception, raise it instead of bare Exception in case of error in ndjson stream. (#2187)
- Preserve volumes order in
client.parse.volumes(...)call. (#2183)
- Expose
neuro_sdk.EnvParseResultandneuro_sdk.VolumeParseResultnames. (#2177)
- Dropped
roleargument fromneuro service-account create: platform automatically generates new role for service account without any permissions. (#2167)
- Add missing requirement to
packaging. (#2160)
-
Add support of
--ownerand--nameoptions toneuro image ls. (#2104) -
Add option
--clusterin commandsneuro ps,neuro topandneuro images. (#2125) -
Removed support of old runtime quota:
neuro config show-quotashows credits and max parallel jobsneuro admin add-user-quota/set-user-quotaonly support credits and max parallel jobs- Added new command
neuro admin get-user-quotato print user quota neuro admin get-cluster-usernow prints table with quota info for each user (#2140)
-
Added
neuro service-accountcommand group. The service account allow to create auth token that can be used for integration with third-party services.neuro service-account create --name optional-name ROLEcreates new service accountneuro service-account lslists service accountsneuro service-account get ID_OR_NAMEretrives single service accountnuero service-account rm ID_OR_NAMEremoves service account and revokes token. (#2147)
-
Ignore files in parent directories are now used when upload a directory to storage or blob storage. (#1901)
-
Add spinners to the some commands that take time to execute. (#2105)
-
All job related commands support now jobs on other clusters. Commands
neuro exec,neuro port-forward,neuro logsandneuro job saveonly support jobs on other clusters if they are specified by URI (jobs on the current cluster can be also specified by bare ID or name). (#2116)
-
Provide forward compatibility with click 8.0 (#2126)
-
Fix hints for 'neuro run' command. (#2127)
-
Fix compatibility with click==8.0. (#2126)
-
Commands
neuro psandneuro topsupport now outputting several fields in one table cell. By default fields ID and NAME and fields STATUS and WHEN are merged. This saves horisontal space and improves output of long fields (e.g. COMMAND, DESCRIPTION or TAGS). (#2062) -
Added automatic re-attach if container is still alive after
neuro run/neuro attachdisconnects. (#2085) -
Added warning about actual privilege level if
neuro grantgrants privilege less then user already has. (#2092)
- Fixed unexpected behavior in
neuro admin update-resource-presetwhen working with existing preset: Now there are two commands:neuro admin add-resource-presetto add new preset, options use default values if not specifiedneuro admin update-resource-presetto update existing preset, options use previous values if not specified Also droppedcluster_nameargument to fix bug when presets were overwritten by active cluster presets. (#2091)
- Added option
--shareinneuro job run. It allows to share a created job with specified user or role. (#2079)
- Fixed handling jobs with
--pass-configinneuro job generate-run-command. (#2074)
- Bump pinned rich version to >=10.0.1 (#2072)
-
Allow bumping jobs life span using
neuro jobs bump-life-span <timedelta>. (#2020) -
Make a job status reason always printed by neuro status if it is available. (#2047)
-
Add support of automatic re-login after config errors. This allows to upgrade to a new CLI version without manual re-login. (#2048)
-
Made more precise datetime formatting for neuro status: now hours and minutes are always printed. Seconds are printed for event that happened in the last hour. (#2053)
-
Storage operations support now URIs with cluster names different from the current cluster name. You can now list or copy to/from other cluster's storage without switching the current cluster name. (#2054)
-
Added printing of ids of removed disk in
neuro disk rm. (#2056) -
Allow to specify date and time in --since/--until relatively to current moment. For example,
neuro ps --since 1dwill show jobs that were created during last day. (#2059) -
Image operations support now URIs with cluster names different from the current cluster name. (#2066)
-
Added hint about headless login when
neuro loginexecutes in non-GUI environment. (#2068)
-
Do not add implicit '\n's to
neuro job generate_run_command <job-id>command output. (#2034) -
Fix bad formatting in
neuro runcommand. (#2055) -
Do not kill a non-tty job when running in non-tty mode automatically. This allows to exit from
neuro run ... | teeusing^Cwithout killing a job. (#2063)
-
Added option
--formatinneuro topfor specifying the output columns. It is similar to option--formatofneuro ps, but support several additional columns. (#2000) -
Added filtering options in
neuro top:--owner,--name,--tag,--description,--sinceand--until. They are similar to corresponding options inneuro ps. By default shown only jobs of the current user, specify--owner=ALLto show jobs of all users. (#2001) -
Add option
--sortinneuro top. (#2002) -
Drop cluster_name and description from default output of
neuro pscommand. (#2009) -
Allow deletion images without tag. Support multiple arguments for
neuro image rmcommand. (#2010)
- Support job statuses "suspended" and "unknown" in
neuro psandneuro top. (#2011)
- Fixed parsing of huge job payloads. (#2004)
-
Added support of named disks:
- create new disk by
neuro disk create --name <disk-name> STORAGEcommand - name can be used to get/delete disk:
neuro disk get <disk-name>orneuro disk delete <disk-name> - name can be used to mount disk:
neuro run -v disk:<disk-name>:/mnt/disk ...(#1983)
- create new disk by
-
Added printing of job id when
neuro floworneuro attachexits. (#1993) -
Added "Life span ends" field to
neuro status <job-id>. (#1994) -
Now all datetime fields are printed in "humanized form". Use
neuro --iso-datetime-format <command> ...to force ISO 8601 format. (#1995) -
Added recovering of attachment to job after internet connection issue in
neuro attachandneuro run. (#1996)
-
neuro topsupports now multiple jobs. If multiple jobs is specified at command line, it outputs a table for that jobs. If no jobs are specified, it outputs all active user's job. (#418) -
Added private option
--x-trace-allwhich forces distribute tracing. (#1973) -
Add
neuro job generate-run-command <job-id-or-name>to simplify reruning of jobs. (#1977) -
Added new options to
neuro ps:--distinct: Show only first job if names are same.--recent-first/--recent-last: Show newer jobs first or last (#1981)
- Soft-deprecate
neuro job tagscommand. (#1840)
-
Removed support of deprecated
--status=allinneuro ps. Use--allinstead. (#1883) -
Refresh config after resource presets are updated. (#1899)
-
Don't open browser on
login-with-token. (#1748) -
Suppress non-critical neuro configuration database errors. (#1816)
-
Fixed unintentional interpretation of sequences like
:+1and[b]in descriptions, commands, paths, URIs, etc when output on the console. (#1917)
- Fix unstable imports of
richtool. (#1911)
- Extract
neuromation.apiintoneuro-sdkpackage, renameneuromationtoneuro-cli. (#1892)
-
Generating the URI string for
RemoteImagenow correctly escapes special characters like "#", "%", "?" in image name and tag. (#1895) -
Fixed conflict between logging and showing progress in
neuro cp -vv. Use therichlibrary for outputting colored logs. (#1897)
-
Added --privileged flag to
neuro run. Added corresponding argumentprivilegedtojobs.start. (#1879) -
Added Disk life-span information to the
neuro disk ls --long-formatandneuro disk getcommands. (#1880)
- Added support of filtering by statuses "cancelled" and "suspended" in
neuro ps. (#1881)
-
Add
-loption toneuro image tagsfor long output format (#1855) -
Add
-fflag toneuro image rmto force delete images that have multiple tag references (#1828) -
Enable display of non-zero restarts count for all jobs (previously only jobs with proper --restart-policy had this field). (#1859)
-
Added cluster resource presets management commands. Added
Preemptible Nodecolumn in resource presets format inneuro config show. (#1863) -
Pass preset instead of container resources to job api. (#1864)
-
neuro blob cpuses now partial read when retry to download a blob. (#1865) -
Change jobs capacity API endpoint. (#1872)
-
Add disk column to
admin get-clusterscommand. (#1873) -
Added options
--updateand--continuefor commandneuro blob cp. Added optional boolean parametersupdateandcontinue_to the corresponding API. (#1875)
- Fixed parsing file modification time in Blob Storage. (#1874)
-
Moved pass_config option to API. Now it uses config completely encoded into ENV variable. (#1814)
-
Added support of max running jobs quota. The
neuro admin set-user-quota -j <count>command configures this quota for user. By default, a new job cannot be created after the quota is reached, but the--wait-for-seatflag allows creating a job that will wait for another job to stop. (#1827) -
When the connection is lost during transferring files
neuro cpretries now only sending and retrieving new data instead of starting file operation from the start. Added optional parameters offset and size in methodstorage.open()for partial reading. Added methodstorage.write()for overwriting the part of the file. (#1831) -
Added option
--continuefor commandneuro cp. It specified, copy only the part of the source file past the end of the destination file and append it to the destination file if the destination file is newer and not longer than the source file. Otherwise copy the whole file. Added corresponding keyword-only boolean parametercontinue_to the API. (#1841) -
Add support of
SUSPENDEDjob status. (#1844) -
Added support of
neuro ps --owner ME. This allows to filter current users jobs. (#1845) -
The active cluster name can now be specified in local configuration file. (#1846)
-
Add
neuro image rmcommand for removing images from remote registries (#1770) -
Allowed to share disks. (#1811)
-
Added support of multiple disk removal:
neuro disk rm disk-id-1 disk-id-2works properly now. (#1821) -
Added displaying of restarts count for restartable jobs. Added
restartsfield toJobStatusHistory. (#1822)
- Bump
aiohttpto >= 3.7.2 in the library requirements, drop transient dependenciesmultidictandyarl. Support Python 3.8 and Python 3.9 (#1802)
- Allowed to share secrets. (#1791)
- Made
neuro completion patchidempotent. (#1760)
- Suppress connection errors when cluster is not available in
neuro config show. (#1763)
-
Fragment, query, user, password and port parts are not allowed now in URIs (these parts were silently ignored before). Fixed support of local file paths containing characters like "#", "%", ":", "?", "@" and "~". (#1531)
-
Implemented disks management commands. (#1716)
-
neuro runallows now to specify disk volumes using-v disk:<DISK>:<PATH>:<RW_FLAG>. (#1721) -
Added support of
created_atandlast_usagefield of disks. Added--long-formatoption toneuro disk ls. (#1729) -
Memory amount parsing now supports of both
bandBsuffixes for specifying decimal quantities. Improvedneuro disk createdocs. (#1731) -
Added a
--life-spanargument toneuro disk create. Added alife_spanargument inclient.disks.create(). (#1734) -
Added output to
neuro disk rmandneuro secret rmwhen-v/--verboseglobal flag is set. (#1738) -
Quite mode command
neuro -q disk lsnow prints simple list of disk ids. (#1740) -
Command
neuro image tagsoutputs now a list of tags instead of a list of images. (#1741) -
Added support of
--full-uritoneuro disk get. (#1747) -
Show a deprecation warning for
--life-span=0inneuro runcommand. (#1749) -
Change the color scheme for job statuses. The yellow color is for the cancellation, cyan is for pending jobs. (#1752)
-
Commands
neuro image lsandneuro image tagsand corresponding APIimages.ls()andimages.tags()can now return more than 100 items. (#1606) -
Make
neuromation.api.SecretFileclass public (#1714)
-
Added support of
CANCELEDstate. (#1696) -
Added support of error messages in streamed delete response. (#1697)
-
Colorize
neuro psoutput. (#1698) -
Added way to destroy browser session by during
neuro logout. (#1699) -
Remove
--volume=ALLoption from CLI and move volume and env variable parsing from CLI module to theParserclass in API (#1707) -
Command
neuro acl listaccepts now a URI prefix for filtering. API functionsusers.get_acl()andusers.get_shares()accept now the uri argument. (#1708) -
neuro runsupports now multiple--env-fileoptions. (#1710)
-
Show add available jobs counts (cluster capacity) in
neuro config showcommand. (#1687) -
Make JobStatus calculation forward-compatible;
JobStatus.UNKNOWNis returned for unknown statuses but the code doesn't raiseValueErrorat least. (#1688)
- Pin
yarlversion dependency to 1.5.1+.
-
Implement
FileStatus.uriproperty. (#1648) -
Add support of plugin-defined config parameters (#1657)
-
Added
find_project_rootfunction (#1660) -
Added
neuro rm --progressandprogressargument toStorage.rmfor tracking delete progress (#1664) -
Added
internal_hostname_namedtoJobDescriptionand to output ofneuro job status. (#1675) -
Added logging of
X-Errorerror description inneuro port-forward. (#1676) -
Added printing of documentation link to
nuero logincommand (#1680) -
Added
neuro admin show-cluster-optionscommand for displaying possible cluster configuration options. (#1681)
-
neuro pssupports now columns "created", "started" and "finished" (hidden by default). (#1020) -
neuro statusshows now the job's http port. (#1375) -
Long list of tags for
neuro statusis now wrapped. (#1408) -
neuro pssupports now the "life_span" column (hidden by default). (#1448) -
Command aliases are now supported if not logged in. (#1480)
-
Added support of the
--schedule-timeoutoption inneuro run. (#1499) -
Added commands for adding and removing roles:
neuro acl add-roleandneuro acl remove-role. (#1582) -
Support expansion of the user home directory ("~") in the file path argument in
neuro secret add. (#1610) -
Put
--helpoption first in the help output for a command or command group (#1627) -
Officially support Python 3.8. (#1638)
-
Create a topic about secrets management and usage. (#1640)
-
Fixed "division by zero" error when copy from
/dev/stdinto storage. (#1129) -
Fixed support of local images whose name starts with Neuro registry prefix, e.g.
registry.neu.ro/alice/imagename. (#1159) -
Fixed support of relative image URIs with numeric name, like
image:5000. (#1631) -
When resolve job URI with other user's name and job name (like
job:/bob/test-name), the owner name no longer ignored when the shared job is not found. (#1633)
- Removed support of
--volume=HOMEinneuro run. (#1202)
-
Implement secrets management commands. (#1545)
-
neuro runallows now to specify secrets: either as a file-v secret:<KEY>:<PATH>or as an environment variable-e <NAME>=secret:<KEY>. (#1558) -
Support azure cluster config file generation. (#1577)
-
Implement
--port-forwardoption forrunandattachcommands. (#1601) -
Take Secrets Service URL from
/config. (#1607)
- Deprecate and hide
submitcommand. (#1602)
-
Always set "LESS=-R" env variable to fix outputs with scrolling, e.g. "neuro help" (#1595)
-
Fix a warning raised by
neuro job port-forwardcommand on Python 3.6 (#1592)
- Increase timeout for waiting for jobs exit after finishing attached session from 15 seconds to 15 minutes. (#1584)
-
A
.neuroignorefile specifies files thatneuro cpshould ignore when upload directories to the storage or object storage. Syntax and semantic are similar to.gitignorefiles. (#1446) -
API:
Client.jobs.run()supports now relative storage URIs for volumes. (#1464) -
Explicit options
--excludeand--includeno longer disable the defaults from the storage.cp-exclude configuration variable. Use explicit--exclude="*" or --include="*"to override default patterns. (#1489) -
Support leading and trailing slashes (
/) in filters. Leading slash prevents matching in subdirectories, e.g./*.txtmatchesspam.txt, but notdir/spam.txt, while*.txtmatches both of them. Patterns with trailing slash match only directories, while patterns without trailing slash match both directories and files. (#1490) -
Implement attach/exec/interactive-run using WebSockets. (#1497)
-
Implement port forwarding over WebSockets. (#1555)
-
Added the
--exclude-from-filesoption inneuro cp.API: Added the optional parameter
ignore_file_namesin theupload_dir()methods ofstorageandblob_storage. (#1560)
- Fix swallowing underscores in
neuro status. (#1565)
- Drop deprected and hidden
neuro storage loadcommand. (#1542)
- Suppress annoying message about improperly handled GeneratorExit exception. (#1525)
-
API:
Storage.ls()is an asynchronous generator now. (#1457) -
Added the
--restartoption toneuro runandneuro submitand API. (#1459) -
API:
Jobs.list()is an asynchronous generator now. (#1473) -
neuro psnow outputs local date instead of UTC and interprets--sinceand--untiloptions as local time if the timezone is not specified. (#1477) -
autocomplete job name/id (#1485)
-
Unencodable characters are now replaced with
?orU+FFFEwhen output to the stdout. (#1502)
-
Fixed downloading a file when it is restarted for some reasons. The newly downloaded data no longer appended to a previously downloaded data, but overwrites it. (#1521)
-
Fixed parsing image URIs with the single slash after scheme, like
image:/otheruser/imagename. (#1505)
- Added
--sinceand--untiloptions toneuro ps. (#1461)
- Make exit code of
job runcommand more reliable. (#1470)
-
Commands
neuro ps,neuro job status,nauro image listandneuro acl listoutput now URIs in the short form if possible. Use the new--full-urioption to get full qualified URIs. (#1330) -
Changed interpretatation of cluster related URIs (with schemes
storage:,image:andjob:) with missed host and path started with/.storage:/user/pathis expanded now tostorage://{defaultcluster}/user/path, so you do not need to specify the cluster name when refer to other user's resources on the same cluster. (#1424) -
Implement
neuro storage treecommand for displaying the directory tree on storage. (#1435) -
Filter patterns are now more compatible with
.gitignore. Added support of**which matches zero or more path components.?and*no longer match/. Patterns which does not contain/at the beginning or middle matches now files in subdirectories. (#1444)
- Bump
typing_extensionsdependency version to satisfy CLI requirements. (#1421)
- Fix
--pass-configerror: File exists: '/root/.neuro'. (#1415)
-
Support job run-time limit:
neuro run --life-span "1d2h3m". (#1325) -
For cluster-specific resources (with schemes
storage,imageandjob) client now use URIs containing the cluster name, e.g.storage://{clustername}/{username}/{path}instead ofstorage://{username}/{path}. Relative URIstorage:{path}will be expanded to absolute URI containing the current cluster name and the user name:storage://{clustername}/{username}/{path}. Same forimageandjobschemes. (#1341) -
Added 'neuro image ls -l' option, which also prints Docker URL-s (#1354)
-
Added the
-uoption inneuro acl listto specify a role or user to which resources are available and for which they are shared. (#1355) -
Sort
neuro admin get-cluster-usersby name. (#1359) -
Steal config files on
neuro job --pass-config. (#1361) -
Support hidden files on the storage. Hide names started with a dot by default, provide
neuro ls --alloption to show all files. (#1362) -
Add HTTP tracing of neuro login commands. (#1387)
-
Support job tags:
neuro run --tag=experiment-1,neuro ps --tag=experiment-1. (#1393) -
Support job tags listing:
neuro job tags. (#1396) -
Optionally display job's tags in
neuro ps(the feature needs to be explicitly enabled in.neuro.tomlconfig file). (#1406)
-
Fix the alias finding routine when user is not logged in. (#1360)
-
neuro killexits now non-zero code if it failed to kill any job in the list. (#1272) -
Support un-quoted commands for neuro-exec:
neuro exec bash -c "ls && pwd"is now a valid syntax. (#1321)
-
Support custom columns format for
neuro pscommand. (#1288) -
Support custom aliases. (#1320)
-
Removed support of
~in URIs (likestorage://~/path/to). Relative URIs can be used instead (likestorage:path/to). Support of tilde in local file paths (like~/path/to) has been preserved. (#1329)
-
--env-filenow allows blank lines and comments (lines starting with "#") in the file. (#1208) -
Send the usage statistics to Google Analytics (#1286)
-
Use Sqlite for saving the
~/.neuro/dbconfiguration file. (#1298)
-
Fix columns width and aligning for
neurp ps --format=<>command. (#1302) -
Fix ps-format documentation issues. (#1303)
-
Parameters specified in ps-format now always override default values.
widthtakes priority overminandmax,maxtakes priority overmin. (#1310)
-
Add
neuro admin add-user-quotaandneuro admin set-user-quotacommands to control user quotas (#1142) -
Added options
--excludeand--includeinneuro storage cp. (#1182) -
Adjust NAME column of
neuro psso that grep by name works. (#1189) -
Read
neuro ps --formatspec from config files if present. (#1294) -
Read
neuro cpfilters from user configuration file. (#1295)
-
Now
neuro imagesdo not require the installed Docker. (#1071) -
Implement
neuro config get-clusterscommand. (#1177) -
Convert configuration file into configuration directory, now it is
~/.neurofolder. (#1183) -
Implement
neuro config switch-clusterfor switching between available clusters. (#1217) -
Implement
neuro admin get-clusterscommand. (#1223) -
Implement
neuro admin add-clustercommand. (#1224) -
Add
neuro admin generate-cluster-configcommand. (#1227) -
neuro project initnow supports argument to set default value for generated project directory (#1230)
- Correctly process both quoted command arguments (
neuro run python:latest "python3 -c 'import os'") as well as unquoted version (neuro run python:latest python3 -c 'import os'`). (#1229)
-
Trace sent HTTP requests and received replies to stderr if
--traceflag is passed. (#467) -
Display
Clusterfield for job status and listing commands. (#874) -
Display
Entrypointfield for job status command. (#924) -
Display volumes information for
neuro statuscommand. (#1003) -
Option
--volume=HOMEdeprecated. (#1009) -
Provide client.presets property, update docs. (#1078)
-
Retry storage operations in case of some errors. (#1107)
-
neuro killwill continue work if multiple jobs specified but you haven't required permissions for some of them. (#1122) -
Introduce
neuro config show-quota. (#1141) -
Use pager for long lists. (#1152)
-
Add global option '--hide-token/--no-hide-token' to be used together with '--trace' for preventing the user's token from being printed to stderr for safety reasons. (#1158)
-
Suppress security checks for config files if NEUROMATION_TRUSTED_CONFIG_PATH environment variable is on. (#1173)
-
When running a job with
--detachoptionneuronow returns an error status if job fails to start at all (e.g., when cluster scale up fails). If job starts successfully (regardless of its run result)neuro runwith--detachreturns 0 like before. (#1059) -
Provide default arguments for api.Resources constructor to keep broken backward compatibility. (#1070)
-
Fix help message for
neuro project init: in fact, the command does not accept an argument. (#1080) -
Process 502 Bad Gateway as a separate exception, don't miss it with 400 Bad Request. (#1111)
-
Wait for
ThreadPoolExecutorfinish befor exit from the program. (#1144)
-
Introduce
neuro project initfor scaffolding an empty project. (#1043) -
neuro cp -rnow works with non-directories. (#1053) -
Disable logging of annoying SSL errors. (#1065)
-
Make image commands work on Windows. (#1067)
-
Fix bug with job-name resolution introduced in release 19.8.19: do not print annoying warning messages if more than one job with specified name was found. (#1034)
-
Fix
neuro job savetiming out too early. (#1062)
- Improved displaying a progress of copying many files. (#1034)
- Deprecate
neuro storage loadcommand. (#1028)
-
Due to
certifiupdate delays, show certifi package error not earlier than two weeks after the latest version was released. (#944) -
Support streaming output for
neuro-savecommand. (#946) -
Implement mounting all available volumes:
neuro run --volume=ALL. (#974) -
Uploading/downloading of a directory containing many small files is now 4-10 times faster. (#981)
-
Support Cloud TPU: new cli options
--tpu-type,--tpu-sw-versionadded. (#982) -
Support job operations via job-URI (e.g.,
neuro status job://owner-name/job-name). (#988) -
Support job filtering by owner:
neuro ps -o user-1 --owner=user-2. (#990) -
Added the
--updateoption inneuro storage cp. It makes the command copying the source file only when it is newer than the destination file or when the destination file is missing. (#1007) -
Added the
-d(--directory) option inneuro storage lswhich makes the command to list directories themselves, not their contents. (#1012)
FileStatus.permissionnow isAction, wasstr. (#963)
- Fix regression: restore port-forward functionality. (#979)
-
Change default interactivity option for
neuro exec: by default, interactive--ttywill be used. (#942) -
Use
Optional[datetime]inJobStatusHistory.created_at,.started_at,.finished_atinstead ofstr. (#955) -
neuro storage cpis now up to 2 times faster for a directory with many small files. (#958) -
Refactor port forwarding api to use async context manager for controlling a time of port forwarding process. (#959)
- Support remote images with registry
host:port. (#939)
-
The
-p/--non-preemptibleparameter forneuro runhas been removed in favor in embedding it into presets coming from the server. (#928) -
Show progress for
neuro cpby default. (#933) -
Use dataclasses in image progress API (#935)
-
Add option
ps --all, deprecateps -s all. (#538) -
Forbid mixing arguments and CLI options in
run,submitandexeccommands. Options (parameters starting from dash or double dash, e.g.-nand--name) should prepend arguments (e.g.image:ubuntu:latest). All tailed options belong to executed container command, not to neuro CLI itself. (#927)
-
The behavior of
neuro storage mvis now closer to the behavior of themvcommand. It moves now files inside the target directory if it exists and moves a file under the new name otherwise. Added also options--target-directory(-t) and--no-target-directory(-T). (#203) -
neuro job submitandneuro job runcommand now support--pass-configoption. This option ensures your local neuromation config file is uploaded to your job allowing you to executreneurocommands from the container. (#827) -
Add
neuro image tagscommand to list all tags for a given image (#852) -
Stabilize jobs API (#879)
-
neuro storage cp,neuro storage mvandneuro storage rmsupport now globbing for source arguments. Globbing can be disabled by specifying new option--no-glob. Added commandneuro storage globwhich lists resources matching patterns. (#889)
-
Implemented
neuro job browse. Added the--browseflag toneuro job submitandneuro job run. (#571) -
Added the global
--quietoption, opposite to--verbose. Both options are additive. The--quiteoptions forneuro jobandneuro imageare deprecated now. (#848) -
Drop
neuromation.api.NetworkPortForwardingfor the sake ofneuromation.api.HTTPPort(#861) -
The output of "storage" commands in verbose mode is now more similar to the output of corresponding POSIX command. In particular
neuro -v storage cp -routputs a line for every copied file or directory. (#869) -
The behavior of the
neuro storage cpis now closer to the behavior of thecpcommand. It now copies files inside the target directory if it exists and copies a file under the new name otherwise. Added also options--target-directory(-t) and--no-target-directory(-T). (#870)
- Fix certifi upgrade suggestion text. (#845)
- Make non-preemtible mode default. (#829)
- Improve storage operations stability by supporting sticky sessions. (#832)
- Relax pyyaml version requirement to don't force users to upgrade it if pyyaml was installed by anaconda or another non-pip installer. (#828)
- Don't run version checks if config is not loaded by CLI command. (#826)
-
Changes in
neuro store mkdirbehavior: fails if the directory already exists or parent directories do not exist. Add option--parentsto make parent directories as needed and ignore existing directories. (#131) -
Changes in
neuro store rmbehavior: removes only files by default, add option--recursiveto remove directories. (#354) -
Storage commands (
neuro storage ls,neuro storage cp,neuro storage mv,neuro storage rm) accept now variable number of arguments. This allows to copy or remove several files by one command. (#784) -
Implement
neuro config login-headlesscommand. (#793) -
Changes in interpretation URIs with absolute path and without host name.
storage:///foo/barmeans now the same asstorage://foo/barinstead ofstorage://{currentuser}/foo/bar, andstorage:///can be used for access to the storage root directory. (#808) -
neuro storage cpnow supports copying to/from non-regular files like character devices and named pipes. In particular this allows to output the file to the stdout or get the input from the stdin (/dev/stdoutand/dev/stdinon Linux,CONon Windows). (#813) -
Relax certifi required version. Raise a warning if the package should be updated. (#819)
- Allow to logout even if config file is broken. (#792)
-
Print exposed HTTP for named jobs (#736)
-
Support retrieving server config for authorized users. (#766)
- Implement job run command. (#652)
- Fix image name parser to substitute lastest tag automatically. (#729)
-
New option
--neuromation-configfor using alternative config file location. Environment variableNEUROMATION_CONFIGcan be used as option. New commandneuro config dockerallows to usedocker image push/pullcommands with platform registry. (#381) -
neuro port-forwardcommand now accepts multiple local-remote port pairs in order to forward several ports by a single command. (#632) -
Support job names. (#648)
-
Make progress argument for storage API optional. (#687)
-
Rename neuromation.client to neuromation.api (#688)
-
Implement
neuro config login-with-token TOKEN URLcommand. (#712)
-
Don't allow to submit image names starting with dash. (#526)
-
Respect
--network-timeoutoption inlogsandcpoperations. (#703)
- Remove deprecated functionality:
neuro model,neuro config id_rsaandneuro job submit --sshoption. (#700)
- Change the default API endpoint to
https://staging.neu.ro/api/v1(#666)
- Fix incorrect
--volumeparsing under windows (#635)
- New flag
--http-auth/--no-http-authwhich controls authentication for forwarded HTTP port. Enabled by default. (#604)
-
Increase wait time to finish on POSIX platforms. It dignificantly decreases a chance of error report after CLI command executions. (#597)
-
Fix forward compatibility with platform server. (#599)
-
Close version checker properly in case of error in a command execution. (#586)
-
Pin aiohttp to 3.5+ to satisfy minimal supported CLI version. (#587)
-
Extended '/dev/shm' space (
--extshm) turned on by default (#449) -
Add support for server-side job list filtering (speedup of the
neuro ps) (#474) -
Several UX improvements. (#486)
-
neuro store lsnow supports colored output, LS_COLORS(GNU) or LSCOLORS(BSD) environment variable required. (#487) -
Improve shell completion logic, make shell type argument mandatory. (#506)
-
Add Http URL to neuro job submit output (#527)
-
Added neuro port-forward command to forward a port of a job exposed during job submit to a local one. (#535)
-
Support Windows platform (#548)
-
Check the lastest PyPI neuromation release, suggest to upgrade if PyPI has a newer version. (#308)
-
Changes in
neuro store lsbehavior: display files by columns by default, add option-lfor long output, display one per line for pipes by default. (#427) -
Set up the platform to work with the new platform DNS names. (#495)
-
Use colored mode only if all stdin, stdout, and stderr are tty. (#473)
-
Improved login error reporting (#477)
-
Add
-p/-Pshortcuts for--preemtible/--non-preemtibleforneuro submitcommand. (#458) -
Wait for job start/failure on job submit. (#356)
-
Support
-vas a shortcut for--volumeinneuro submit ...command. (#383) -
Improve resource URI normalization and print normalized URIs back to the user. (#457)
-
Re-organize CLI commands for better UI/UX experience. Obsolete commands are hidden but still supported. (#460)
-
Add top-level aliases for the most frequent commands. (#439)
-
Better formatting for examples section. (#446)
-
Bump
clickversion to7.0+. (#437) -
Temporary disable scary logging about unhandled exception. (#438)
-
Fix an error in local path normalization. (#443)
-
Implement
neuro job top(#412) -
Pretty format output for
neuro config showcommand, print current authentication token byneuro config show-token. (#426) -
Check
~/.nmrcconfig file for strict0o600permissions. (#430)
- Drop
--tokenand--urlparameters from the root command, useneuro loginandneuro config urlto setup config parameters. (#430)
-
Deep refactor
neuro storecommand and corresponding API client. (#324) -
Default API URL switched to HTTPS for neuro cli (#325)
-
Job resource output formatting for command
neuro job statuschanged (#328) -
neuro image pull/pushcommand improved for support different names/tags for images, introduced 'image://' scheme for image referencing. (#349) -
Implement --show-traceback command line option to show python traceback in case of top-level error (#365)
-
Added new option
--insecurefor storing auth in plain text file instead system keyring. (#366) -
New
neuro image lscommand for listing custom images available on platform repository. (#367) -
Added new command
neuro job execto execute commands in already running job without ssh server. (#373) -
Display Preemptible in job status output (#393)
-
Make the client work on Python 3.7 (#402)
-
Implement
neuro job top(#412)
-
Jobs datastructure refactored (#320)
-
Removed keyrings.cryptfile from project dependencies. Please remove it self if upgrade neuromation. (#366)
- Support Python 3.8 and 3.9 (#249)
- Reduce neuro-extras image size, add neuro-flow into the image and include GitHub ssh key to known hosts. (#290)
- Image transfer now works even if it was initiated not from SRC cluster. (#295)
- Make
neuro-extras data cpaccept preset name and migrate code to useneuro.jobs.startmethod in the SDK (#218)
- Downgrade Kaniko to v1.3.0 due to OOMKills (#287)
-
Fix
neuro-extras data transferjob lifespan (set to 10 days), add usage hints, fix seldon init container user ID (#187) -
Fix connection issues in image build jobs by relying on neuro-cli instead of neuro-sdk, add support of job tags for image build jobs. (#205)
No significant changes.
No significant changes.
No significant changes.
-
Reduce verbosity for image building (to turn it back on:
neuro-extras image build --verbose ...). (#84) -
Add
--presetparameter toneuro-extras image buildsubcommand in order to specify preset (#136) -
Add support for HTTP(S) data ingestion (#150)
- Speedup image build by Kaniko downgrade to v1.1.0 (COPY layers caching), replace /workspace context dir to /kaniko_context. (#157)
- Bump neuromation to >=20.10.30 (#141)
-
Fix exit-code for image builder job (#82)
-
Support full storage URI for both SRC and DST in
neuro-extras data transfer(#112) -
Fix data ingestion to/from GCP (move GCS auth to entrypoint) & add hints to CLI commands. (#115)
-
data cp: fix source archive removal if SRC is a local file; add flag for optional TMP folder usage (#118)
-
Fix image transfer: switch cluster properly (#130)
- Add
-c/--compressoption todata cpto support data packing before sending to destination (#38)
No significant changes.
- Change image's entrypoint from
neurotobash. (#81)
- Add
towncrierfor release notes management (#78)
- Renamed
inheritsyaml property tomixins. (#560) - Added support project level
mixins. Mixins defined inproject.ymlavailable both for live and batch configurations, so they cannot define any live or batch specific properties. (#562) - Added command
neuro-flow delete-project <project-id>that allows complete removal of project. (#581)
- Fix crash with global images in batch mode. (#544)
- Added new sections
defaults,images,volumesto theproject.ymlfile. The work the same as the do inlive/batchexcept they are global -- everything defined inproject.ymlapplies to all workflows. (#506)
-
Added modules feature: a module is simillar to action but has following differneces:
- only local files support (
ws:scheme), github located modules are forbidden. - the module content has access to workflow global defaults. The default env/volumes/preset
is inherited, and expressions can access top level contexts such as
flow.
Calling a module is similar to calling an action, except for use of
moduleproperty instead ofactionproperty:Live module call:
jobs: test: module: ws:live-module args: arg1: val 1Batch module call:
tasks: - id: test module: ws:batch-module args: arg1: val 1 ``` ([#464](https://github.com/neuro-inc/neuro-flow/issues/464)) - only local files support (
-
Add support of mixins in live and batch mode.
Mixins allow to define and reuse some common parts of jobs in live mode and tasks in batch mode. To define a mixin, use top level
mixinssections:... mixins: credentials: volumes: - secret:some_key:/var/some_key env: KEY_LOCATION: /var/some_key ...Mixins can define the same properties as "job" and "task" do, except for "id" field in batch mode. To apply a mixin, use
inheritsproperty:... jobs: test_a: inherits: [ credentials ] ...When applied, all fields defined in mixin will be merged into job definition. You can apply multiple mixins simultaneously. If some of applied mixins share the same property, or/and it is defined in task, the following rules are applied:
- If property is scalar (int/string/bool), then job will use value defined in job definition. If it is absent, the value from the rightmost of the mixins that define this property will be used:
mixins: mix1: image: example_mix1 mix2: image: example_mix2 jobs: job1: inherits: [mix1, mix2] image: example job2: inherits: [mix1, mix2]The
job1will useexampleimage, andjob2will useexample_mix2image.- If property is list, that all lists will be concatenated.
- If property is dict, that rule for scalars will be applied for dict values separately.
Mixins can inherit from each other:
mixins: mix1: env: TEST: 1 mix2: inherits: [ mix1 ] image: example ``` ([#498](https://github.com/neuro-inc/neuro-flow/issues/498))
- Improved performance of
neuro flow bakesit will now print only new rows instead of re-rendering all table (#501)
- Implement retry on NDJSON errors (#493)
- Fix cache key calculation: now it doesn't depend on top-level contexts but on a task calculated parameters plus "needs" and "state" for stateful actions only.
- Fix READ retry logic: increase delay between retries (was decreasing to zero in 21.6.17 release by mistake).
-
Added command to clear cache of a single task:
neuro-flow clear-cache <batch> <task_id>. (#452) -
Added support in live jobs params. The following contexts are now available under
jobs.<job-id>.params:project,flow,env,tags,volumes,images. (#457) -
Generate default live job name as '--[<MULTI_SUFFIX>]' if not set. (#462)
-
Retry operations inside a batch runner if possible; it increase baking stability. (#483)
-
Fixed bug when executor ignored bake cancellation. (#449)
-
Fixed wrong filename of action config file in logs. (#450)
-
Fixed missing id of initial job for cached tasks. (#451)
-
Don't attach to live job in dry-run mode. (#463)
- Fixed hanging when executor tried to build an image in batch mode. (#448)
-
Fix broken cache when using images: now remote context dir name is based on image ref instead of random name. (#422)
-
Fix a error that leads to bakes cache check miss even if the record is present in cache actually. (#441)
-
Support shared projects. Shared project should have parameters
ownerorroleset inproject.yml. (#373) -
Add bake id to
neuro-flow inspect. (#396) -
Added support of
storage:urls inimages.<image-id>.contextandimages.<image-id>.dockerfile. (#402) -
Added image building functionality to batch mode. The
imagesyaml section entries withcontextanddockerfileare now allowed both in batch and batch action config files. Image build starts automatically when task that uses it is ready to be run. (#412) -
Added automatic creation of parent directories in
neuro-flow uploadcommand. (#416) -
Added cancellation of image build jobs if bake is cancelled or failed. (#423)
-
Added
force_rebuildflag to image section in yaml config. (#424) -
Added new options to neuro-flow bakes:
--since DATE_OR_TIMEDELTAto show bakes that were created after specified moment--until DATE_OR_TIMEDELTAto show bakes that were created before specified moment--recent-first/--recent-lastto alter ordering in the result table (#428)
-
Pre-fetch the last attempt in bakes list to speed up the command. (#429)
-
Fixed auto-generation of suffixes for multi jobs in live mode. (#415)
-
Fixed overriding param with empty value,
--param name ""works properly now. (#417) -
Fixed EvalError when tried to access
ref,ref_rw,ref_roof volume context. (#418)
- Fix a bug with processing SKIPPED task
-
Neuro Flow now uses the dedicated Platform API to store the flow database. The storage is still supported but will be removed in a few months.
-
Added new expressions functions:
values(dict_instance): get values of dictionary (similar to python'sdict_instance.values())str(any): convert any object to stringreplace(string, old, new): replace all occurrences ofoldinstringwithnew.join(separator, array): concatenate array of strings insertingseparatorin between. (#357)
-
Added support of default volumes similar to default env: both in live and batch modes, you can specify them under
defaultssection:defaults: volumes: - storage:some/dir:/mnt/some/dir - storage:some/another/dir:/mnt/some/another/dir- In live mode such volumes will be added to all jobs.
- In batch mode such volumes will be added to all tasks.
Default volumes are not passed to actions (same as default env). (#359)
-
Added passing of global options (
-v,-q,--show-traceback) to neuro cli and executor. (#360) -
Added
--dry-runflag forneuro-flow runthat enables prints job command instead of running it. (#362) -
Added support of tagging bakes.
To tag a bake:
neuro bake --tag tag1 --tag tag2 batch_nameTo retrieve bakes by tags:
neuro bakes --tag tag1 ``` ([#382](https://github.com/neuro-inc/neuro-flow/issues/382))
-
Fixed bug that led to crash in
neuro-flow inspectwhen bake had cached task. (#358) -
Support click 8.0 (#407)
-
Mark cached task in
neuro-flow inspect <bake>as "cached" instead of "succeeded". (#318) -
Auto create parent directories in "upload()" expression function. (#319)
-
Add bake_id tag to jobs started by bake inside neuro platform. (#320)
-
Add tags to remote executor jobs. The tags are: "project:project_id", "flow:flow_name", "bake:bake_id", "remote_executor". (#321)
-
Print name of the action in error about unexpected needs entry, for example:
ERROR: Action 'ws:some_action' does not contain task 'wrong_task_name' ``` ([#323](https://github.com/neuro-inc/neuro-flow/issues/323)) -
Added printing of filename in expression evaluation errors. (#324)
-
Dropped
outputs.needsin batch actions. Made all action task results available for calculation action needs. This avoids confusing behavior when action can succeed even when some of its tasks have failed. (#325) -
Add support of empty list and dict (
[]and{}) in expressions. (#333) -
Added validation of tasks
needsproperty. (#334) -
Added validation of action arguments before starting a bake. (#336)
-
Implemented marking of bake as failed when an error happens during the batch execution. If the error is caused because of SIGINT, the bake will be marked as cancelled instead of failed. (#338)
-
Allow to specify timeout for executor job in yaml and increase default lifespan to 7d. Example:
kind: batch life_span: 30d tasks: ... ``` ([#339](https://github.com/neuro-inc/neuro-flow/issues/339)) -
Add support of local actions inside of batch actions if they do not depend on any remote tasks. (#340)
- Use 1-based indexes instead of 0-based for lines and columns in error messages. (#335)
- Fix executor restart when there is a action that should be marked as ready. (#315)
- Add support of parsing batch params from file. (#295)
-
Added proper error message for the hosted on GitHub actions, which reference is wrong (URL leads to 404). (#294)
-
Fix processing of bake that has actions that use another bake action. (#302)
- Fixed parsing of needs in actions
- Enable restarting of remote executor jobs on failure.
- Support dependencies on running tasks along with finished ones in batch mode. (#255)
- Fix windows path issue. (#261)
- Implement
inspect_job()expression function. (#255)
- Fix ignoring of workdir in batch mode. (#261)
-
Fixed operator precedence: previously all operators had same precedence. Made
orandandoperate as logical operators instead of bitwise. (#239) -
Forbid passing args into a job, volumes into an action etc. (#241)
-
Allow schedule timeout parameterization in the flow description; useful in cases, when the job should be launched on scarce resources. (#202)
-
Allow image overwrite by forwarding the
--force-overwriteflag to the underlyingneuro-extras image buildcommand. (#203) -
Support of preset parameterization for image build job; now user could change the hardware environment for image build. (#204)
-
Implement
parse_volume()expression function. (#217) -
Support compound expressions for
volumes,tags,env,port_forwardattributes:jobs: job_a: volumes: "${{ ['ubuntu', volumes.volume_a.ref] }}" ``` ([#236](https://github.com/neuro-inc/neuro-flow/issues/236))
- Fix
hash_files()function: it is faster, hashes the whole file instead of the first 256 KiB, and includes processed filenames into the hash. (#190)
- Rework output of
neuro bakecommand; now with timestamps and progress bars. (#172)
-
Added validation of
outputs.needsin batch actions. (#163) -
Store a JSON with live jobs configuration on every run job start. (#170)
-
Implement
lower()andupper()expression functions, e.g.ref: image:${{ lower(flow.project_id) }}:v1.0(#174)
- Don't raise an error in
neuro-flow runif the remote folder for the project already exists (#184)
- Improve inspect command: sort output by task creation time, add columns with task start and task finish times. (#156)
-
Fix variable substitution for printed text for bake cancellation (#152)
-
Add forgotten restart command (#160)
- Ignore hidden files and folders (e.g. .cache) when getting the list of bakes (#146)
-
Fix noisy error report for
neuro kill ALLcommand. (#136) -
Fix invariant by saving both started and finished records for cached task (#139)
-
Added --param option to
neuro-flow bakecommand that allows to pass arguments toparamssection of the config file. (#128) -
Added --param option to
neuro-flow runcommand that allows to pass arguments toparamsection of the job in config file. (#132) -
Added
upload()function to re-upload volumes just before starting job in live mode. (#133)
No significant changes.