Write tests for the app

[Namit2111]

We need tests for the whole app including backend and frontend that we can run on PR before merging them.

[rajatkriplani]

Please refer to my fork before going through this comment. I'm getting some errors while running the test. And the error occurs during the import of app in your test_app.py and originates from the NLTK library when trying to read the 'bible-kjv.txt' file from the Gutenberg corpus. Here's the relevant part of the traceback (from terminal):

backend/utils/utils.py:15: in <module>
    bible = gutenberg.raw('bible-kjv.txt')
myenv/lib/python3.12/site-packages/nltk/corpus/reader/api.py:219: in raw
    contents.append(fp.read())
myenv/lib/python3.12/site-packages/nltk/data.py:1139: in read
    chars = self._read(size)
myenv/lib/python3.12/site-packages/nltk/data.py:1425: in _read
    bytes = self.bytebuffer + new_bytes
E   TypeError: can't concat str to bytes

AND

[nltk_data] Downloading package gutenberg to /home/rajat/nltk_data...
[nltk_data]   Unzipping corpora/gutenberg.zip.
[nltk_data] [Errno 2] No such file or directory:
[nltk_data]     '/home/rajat/nltk_data/corpora/gutenberg.zip'

• NLTK is attempting to load the gutenberg corpus. specifically it tries to load 'bible-kjv.txt' using gutenberg.raw('bible-kjv.txt').
• TypeError occurs in NLTK's internal code it suggests a mismatch between expected data types (bytes vs. strings).
• corrupted or missing data error messages indicate that the Gutenberg corpus might not be properly downloaded or accessible.
• problem is that the code in utils/utils.py is executed immediately when the module is imported, even during testing.
• when test_app.py imports app, which in turn imports utils.utils, the code that loads the corpus is executed, leading to the error.

fixes as follows:

• Refactor utils/utils.py to avoid executing code at dodule level, move code that loads data or models into functions that are called when needed.
• mock NLTK data tests, prevent it from attempting to download or access data during tests.
• And ensure NLTK Data is properly installed.

[JustinhSE]

@Namit2111

[Namit2111]

hi @rajatkriplani which fork are you referring to can you please link it ?

[rajatkriplani]

hi @rajatkriplani which fork are you referring to can you please link it ?

You can check the test_app.py here.

[Namit2111]

hi @rajatkriplani when are you available? ping me in discord

[d-beloved]

@JustinhSE @Namit2111 I think we should break this issue into two, for writing tests for the Frontend and another for Backend tests, then we can consolidate the check via a CI workflow