Replace bundled conda-index with standalone conda-index package #4690
+227
−2,549
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,272 @@ | ||
| # Copyright (C) 2014 Anaconda, Inc | ||
| # SPDX-License-Identifier: BSD-3-Clause | ||
| """ | ||
| conda-build's use of conda-index, delegated to now-separate conda-index package. | ||
| """ | ||
|
|
||
| import conda_index.index | ||
|
|
||
| import json | ||
| import os | ||
| from os.path import ( | ||
| dirname, | ||
| ) | ||
| import sys | ||
| import time | ||
|
|
||
| from functools import partial | ||
| import logging | ||
|
|
||
| from concurrent.futures import Executor | ||
|
|
||
| from conda_build import conda_interface, utils | ||
| from .conda_interface import context | ||
| from .conda_interface import CondaHTTPError, get_index, url_path | ||
| from .utils import ( | ||
| JSONDecodeError, | ||
| get_logger, | ||
| ) | ||
|
|
||
| log = get_logger(__name__) | ||
|
|
||
|
|
||
| # use this for debugging, because ProcessPoolExecutor isn't pdb/ipdb friendly | ||
| class DummyExecutor(Executor): | ||
| def map(self, func, *iterables): | ||
| for iterable in iterables: | ||
| for thing in iterable: | ||
| yield func(thing) | ||
|
|
||
|
|
||
| try: | ||
| from conda.base.constants import NAMESPACES_MAP, NAMESPACE_PACKAGE_NAMES | ||
| except ImportError: | ||
| NAMESPACES_MAP = { # base package name, namespace | ||
| "python": "python", | ||
| "r": "r", | ||
| "r-base": "r", | ||
| "mro-base": "r", | ||
| "mro-base_impl": "r", | ||
| "erlang": "erlang", | ||
| "java": "java", | ||
| "openjdk": "java", | ||
| "julia": "julia", | ||
| "latex": "latex", | ||
| "lua": "lua", | ||
| "nodejs": "js", | ||
| "perl": "perl", | ||
| "php": "php", | ||
| "ruby": "ruby", | ||
| "m2-base": "m2", | ||
| "msys2-conda-epoch": "m2w64", | ||
| } | ||
| NAMESPACE_PACKAGE_NAMES = frozenset(NAMESPACES_MAP) | ||
| NAMESPACES = frozenset(NAMESPACES_MAP.values()) | ||
|
There was a problem hiding this comment. it wouldn't be surprising if conda.base.constants exists in all conda > 4.7 or the oldest one we can depend on. |
||
|
|
||
| local_index_timestamp = 0 | ||
| cached_index = None | ||
| local_subdir = "" | ||
| local_output_folder = "" | ||
| cached_channels = [] | ||
| channel_data = {} | ||
|
|
||
|
|
||
| # TODO: support for libarchive seems to have broken ability to use multiple threads here. | ||
| # The new conda format is so much faster that it more than makes up for it. However, it | ||
| # would be nice to fix this at some point. | ||
| try: | ||
| _os_cpu_count = os.cpu_count() or 1 # can be None in rare cases | ||
| except AttributeError: | ||
| _os_cpu_count = 1 | ||
| MAX_THREADS_DEFAULT = _os_cpu_count | ||
|
|
||
| if ( | ||
| sys.platform == "win32" | ||
| ): # see https://github.com/python/cpython/commit/8ea0fd85bc67438f679491fae29dfe0a3961900a | ||
| MAX_THREADS_DEFAULT = min(48, MAX_THREADS_DEFAULT) | ||
| LOCK_TIMEOUT_SECS = 3 * 3600 | ||
| LOCKFILE_NAME = ".lock" | ||
|
|
||
| # TODO: this is to make sure that the index doesn't leak tokens. It breaks use of private channels, though. | ||
| # os.environ['CONDA_ADD_ANACONDA_TOKEN'] = "false" | ||
|
|
||
|
|
||
| def get_build_index( | ||
| subdir, | ||
| bldpkgs_dir, | ||
| output_folder=None, | ||
| clear_cache=False, | ||
| omit_defaults=False, | ||
| channel_urls=None, | ||
| debug=False, | ||
| verbose=True, | ||
| locking=None, | ||
| timeout=None, | ||
| ): | ||
| global local_index_timestamp | ||
| global local_subdir | ||
| global local_output_folder | ||
| global cached_index | ||
| global cached_channels | ||
| global channel_data | ||
| mtime = 0 | ||
|
|
||
| channel_urls = list(utils.ensure_list(channel_urls)) | ||
|
|
||
| if not output_folder: | ||
| output_folder = dirname(bldpkgs_dir) | ||
|
|
||
| # check file modification time - this is the age of our local index. | ||
| index_file = os.path.join(output_folder, subdir, "repodata.json") | ||
| if os.path.isfile(index_file): | ||
| mtime = os.path.getmtime(index_file) | ||
|
|
||
| if ( | ||
| clear_cache | ||
| or not os.path.isfile(index_file) | ||
| or local_subdir != subdir | ||
| or local_output_folder != output_folder | ||
| or mtime > local_index_timestamp | ||
| or cached_channels != channel_urls | ||
| ): | ||
|
|
||
| # priority: (local as either croot or output_folder IF NOT EXPLICITLY IN CHANNEL ARGS), | ||
| # then channels passed as args (if local in this, it remains in same order), | ||
| # then channels from condarc. | ||
| urls = list(channel_urls) | ||
|
|
||
| loggers = utils.LoggingContext.default_loggers + [__name__] | ||
| if debug: | ||
| log_context = partial(utils.LoggingContext, logging.DEBUG, loggers=loggers) | ||
| elif verbose: | ||
| log_context = partial(utils.LoggingContext, logging.WARN, loggers=loggers) | ||
| else: | ||
| log_context = partial( | ||
| utils.LoggingContext, logging.CRITICAL + 1, loggers=loggers | ||
| ) | ||
| with log_context(): | ||
| # this is where we add the "local" channel. It's a little smarter than conda, because | ||
| # conda does not know about our output_folder when it is not the default setting. | ||
| if os.path.isdir(output_folder): | ||
| local_path = url_path(output_folder) | ||
| # replace local with the appropriate real channel. Order is maintained. | ||
| urls = [url if url != "local" else local_path for url in urls] | ||
| if local_path not in urls: | ||
| urls.insert(0, local_path) | ||
| _ensure_valid_channel(output_folder, subdir) | ||
| update_index(output_folder, verbose=debug) | ||
|
|
||
| # replace noarch with native subdir - this ends up building an index with both the | ||
| # native content and the noarch content. | ||
|
|
||
| if subdir == "noarch": | ||
| subdir = conda_interface.subdir | ||
| try: | ||
| cached_index = get_index( | ||
| channel_urls=urls, | ||
| prepend=not omit_defaults, | ||
| use_local=False, | ||
| use_cache=context.offline, | ||
| platform=subdir, | ||
| ) | ||
| # HACK: defaults does not have the many subfolders we support. Omit it and | ||
| # try again. | ||
| except CondaHTTPError: | ||
| if "defaults" in urls: | ||
| urls.remove("defaults") | ||
| cached_index = get_index( | ||
| channel_urls=urls, | ||
| prepend=omit_defaults, | ||
| use_local=False, | ||
| use_cache=context.offline, | ||
| platform=subdir, | ||
| ) | ||
|
|
||
| expanded_channels = {rec.channel for rec in cached_index.values()} | ||
|
|
||
| superchannel = {} | ||
| # we need channeldata.json too, as it is a more reliable source of run_exports data | ||
| for channel in expanded_channels: | ||
| if channel.scheme == "file": | ||
| location = channel.location | ||
| if utils.on_win: | ||
| location = location.lstrip("/") | ||
| elif not os.path.isabs(channel.location) and os.path.exists( | ||
| os.path.join(os.path.sep, channel.location) | ||
| ): | ||
| location = os.path.join(os.path.sep, channel.location) | ||
| channeldata_file = os.path.join( | ||
| location, channel.name, "channeldata.json" | ||
| ) | ||
| retry = 0 | ||
| max_retries = 1 | ||
| if os.path.isfile(channeldata_file): | ||
| while retry < max_retries: | ||
| try: | ||
| with open(channeldata_file, "r+") as f: | ||
| channel_data[channel.name] = json.load(f) | ||
| break | ||
| except (OSError, JSONDecodeError): | ||
| time.sleep(0.2) | ||
| retry += 1 | ||
| else: | ||
| # download channeldata.json for url | ||
| if not context.offline: | ||
| try: | ||
| channel_data[channel.name] = utils.download_channeldata( | ||
| channel.base_url + "/channeldata.json" | ||
| ) | ||
| except CondaHTTPError: | ||
| continue | ||
| # collapse defaults metachannel back into one superchannel, merging channeldata | ||
| if channel.base_url in context.default_channels and channel_data.get( | ||
| channel.name | ||
| ): | ||
| packages = superchannel.get("packages", {}) | ||
| packages.update(channel_data[channel.name]) | ||
| superchannel["packages"] = packages | ||
| channel_data["defaults"] = superchannel | ||
| local_index_timestamp = os.path.getmtime(index_file) | ||
| local_subdir = subdir | ||
| local_output_folder = output_folder | ||
| cached_channels = channel_urls | ||
| return cached_index, local_index_timestamp, channel_data | ||
|
|
||
|
|
||
| def _ensure_valid_channel(local_folder, subdir): | ||
| for folder in {subdir, "noarch"}: | ||
| path = os.path.join(local_folder, folder) | ||
| if not os.path.isdir(path): | ||
| os.makedirs(path) | ||
|
|
||
|
|
||
| def update_index( | ||
| dir_path, | ||
| check_md5=False, | ||
| channel_name=None, | ||
| patch_generator=None, | ||
| threads=MAX_THREADS_DEFAULT, | ||
| verbose=False, | ||
| progress=False, | ||
| hotfix_source_repo=None, | ||
| subdirs=None, | ||
| warn=True, | ||
| current_index_versions=None, | ||
| debug=False, | ||
| index_file=None, | ||
| ): | ||
| return conda_index.index.update_index( | ||
| dir_path, | ||
| check_md5=check_md5, | ||
| channel_name=channel_name, | ||
| patch_generator=patch_generator, | ||
| threads=threads, | ||
| verbose=verbose, | ||
| progress=progress, | ||
| # hotfix_source_repo=None, # unused | ||
| subdirs=subdirs, | ||
| warn=warn, | ||
| current_index_versions=current_index_versions, | ||
| debug=debug, | ||
| # index_file=None, # unused | ||
| ) | ||
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Copied to make keeping track of all the changes easier, and to discourage future
conda_build.indeximportsThere was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
To clarify, we couldn't continue to have index.py as the place for this code to live? Or asked differently, the get_build_index function, is that something that we have in conda-index? I'm a little worried we're introducing another place where users will end up importing from :)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
get_build_indexis only in conda-build. It is mostly about fetching repodata.json instead of generating repodata.json. It would of course be possible to keep it inconda_build.indexThere was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
However we might want to raise deprecationwarning when conda_build.index is imported.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I don't think this submodule should exist.
Plus, all of
conda.indexshould be deprecated with warnings to import fromconda-indexinstead.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Do you mean
conda.indexorconda_build.indexReminder that conda-build's
get_build_indexdoesn't belong in conda-index.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@kenodegard @jezdez I moved this module back to conda_build.index