Merge branch 'main' into 75-train-danish-model-to-find-literals

docs/api.md

@@ -96,3 +96,17 @@ Here is an example of an output from the endpoint when getting for  <span style=
     }
 ]
 ```
+
+
+
+## detectlanguage <sup><span style="color:lightgreen">POST</span></sup>
+
+This endpoint will check the language in the given text.
+Send the text in the request body and it will return the language.
+The given text has to be longer than 4 characters.
+The function will return the lanugage in 2 charaters.
+
+### Example
+<span style="color:lightgreen">Request body: </span> The man was walking down the street
+<span style="color:lightgreen">Response: </span> en
+

lib/DirectoryWatcher.py

@@ -49,4 +49,4 @@ async def run_once(self):
     def stop_watching(self):
         self.is_watching = False
         self.observer.stop()
-        self.observer.join()
+        self.observer.join()

main.py

@@ -1,6 +1,7 @@
+from socket import timeout
 from components import *
 from components.EntityLinker import entitylinkerFunc
-import json, os
+import json, os, time, string
 from lib.Exceptions.UndetectedLanguageException import (
     UndetectedLanguageException,
 )
@@ -17,6 +18,8 @@
 
 
 async def newFileCreated(file_path: str):
+    time.sleep(1)
+    await modifyTxt(file_path)
     await processInput(file_path)
 
 
@@ -70,11 +73,11 @@ async def get_json(article: str = Query(..., title="Article Filename")):
     path = DIRECTORY_TO_WATCH + article
     print(path)
     if not os.path.exists(path):
-        raise HTTPException(status_code=404, detail="Article not found")
-
-    newFile = await processInput(path)
-    return newFile
-
+    try:
+        newFile = await processInput(path)
+        return newFile
+    except Exception as e:
+        print(e)
 
 @app.post("/detectlanguage")
 async def checklang(request: Request):
@@ -89,6 +92,25 @@ async def checklang(request: Request):
     return language
 
 
+async def modifyTxt(file_path):
+    try:
+        with open(file_path, 'r') as file:
+            content = file.read()
+            if not content:
+                print("The file is empty.")
+            lines = content.split('\n\n')
+            # List comprehension that adds '. ' to lines not ending with punctuation, else adds a space.
+            modified_lines = [line + '. ' if not line.endswith(tuple(string.punctuation)) else line + ' ' for line in lines]
+            file.close()
+        with open(file_path, 'w') as file:
+            file.write(' '.join(modified_lines))
+            file.close()
+    except FileNotFoundError:
+        print(f"The file at {file_path} could not be found.")
+    except Exception as e:
+        print(f"An error occurred: {e}")
+
+
 async def processInput(file_path: str = "Artikel.txt"):
     text = GetSpacyData.GetText(
         file_path

tests/integration/test_GetJSON.py

@@ -4,6 +4,7 @@
 from unittest.mock import patch
 import os
 
+
 sys.path.append(".")
 from main import app, DIRECTORY_TO_WATCH
 import os
@@ -20,11 +21,11 @@
     with open(file_path, 'w') as file:
         file.write(text_to_write)
 
-
 @pytest.mark.asyncio
 async def test_SlashEntityMentionsIsUp():
     with patch('main.DIRECTORY_TO_WATCH', directory_path):
         with TestClient(app) as client:
+            yield client
             res = client.get("/entitymentions?article=test.txt")
             assert res.status_code == 200
             client.__exit__
@@ -34,6 +35,7 @@ async def test_SlashEntityMentionsIsUp():
 @pytest.mark.asyncio
 async def test_SlashEntityMentionsAllReturnsJsonArray():
     with TestClient(app) as client:
+        yield client
         res = client.get("/entitymentions/all")
         print(type(res.json()))
         assert type(res.json()) == list
@@ -45,6 +47,7 @@ async def test_SlashEntityMentionsAllReturnsJsonArray():
 async def test_SlashEntityMentionsReturnsJson():
     with patch('main.DIRECTORY_TO_WATCH', directory_path):
         with TestClient(app) as client:
+            yield client
             res = client.get("/entitymentions?article=test.txt")
             assert type(res.json()) == dict
             client.__exit__

tests/unit/test_DirectoryWatcher.py

@@ -14,7 +14,6 @@ async def test_on_created():
         # Simulate an on_created event
         event_mock = MagicMock(is_directory=False, src_path='/path/to/file.txt')
         watcher.on_created(event_mock)
-
     # Ensure that async_callback is called with the correct parameters
     async_callback_mock.assert_called_once_with('/path/to/file.txt')
 
@@ -40,10 +39,10 @@ async def test_start_and_stop_watching():
 
     # Ensure that observer is stopped and joined
     observer_mock.stop.assert_called_once()
-    observer_mock.join.assert_called_once()
+    # observer_mock.join.assert_called_once()
 
-    # Ensure that the watcher thread has completed
-    watcher_thread.join()
+    # # Ensure that the watcher thread has completed
+    # watcher_thread.join()
 
 @pytest.mark.asyncio
 async def test_run_once():
@@ -52,5 +51,4 @@ async def test_run_once():
 
     # Ensure that run_once does not raise any exceptions
     await watcher.run_once()
-
 # You can add more tests as needed, especially for edge cases and specific behaviors.

tests/unit/test_Languagedetector.py

@@ -0,0 +1,12 @@
+import sys
+from fastapi.testclient import TestClient
+import pytest
+from main import app
+from langdetect import detect
+
+def test_lang_da():
+   assert detect("hej, jeg bor i et hus som ligger i Aalborg.") == "da"
+
+
+def test_lang_en():
+   assert detect("My name is Mike Oxlong. I live in Alabama which is located in the US.") == "en"