Fix/deployment script

[pmalarme]

This pull request includes multiple changes across different files to improve the deployment process, update dependencies, and enhance the functionality of the PDF indexer. Below are the most important changes grouped by theme:

Deployment Improvements:

• infra/deploy.sh: Introduced a RANDOM_STRING variable to ensure unique names for the storage account, and updated the method to retrieve the Cosmos DB connection string. [1] [2]

PDF Indexer Enhancements:

• src/pdf_indexer/pdf_indexer.py: Added a document_id to each document's metadata instead of an id, included retry options for the Azure Search client, and implemented batching for adding documents to the vector store. [1] [2]

Dependency Updates:

• src/pdf_indexer/requirements.txt: Updated several dependencies including langchain-text-splitters, langchain-openai, langchain-community, pypdf, tiktoken, and added azure-search-documents and azure-identity.
• src/podcast_generator/requirements.txt: Updated dependencies such as openai, langchain-openai, langchain-community, and langchain-core.

Miscellaneous Changes:

• .vscode/settings.json: Added the word "Podcaster" to the spell checker word list.
• src/pdf_indexer/pdf_indexer.py: Imported the time module.
• src/podcast_generator/podcast_generator.py: Added an extra newline for code readability.

[Copilot]

Copilot reviewed 3 out of 7 changed files in this pull request and generated no comments.

Files not reviewed (4)

• .vscode/settings.json: Language not supported
• infra/deploy.sh: Language not supported
• src/pdf_indexer/requirements.txt: Language not supported
• src/podcast_generator/requirements.txt: Language not supported

Comments suppressed due to low confidence (2)

src/pdf_indexer/pdf_indexer.py:166

• [nitpick] The variable name 'splits_batch' is ambiguous. It should be renamed to 'document_batch' for better clarity.

splits_batch = splits[i:i+batch_size]

src/pdf_indexer/pdf_indexer.py:167

• Replace the print statement with a proper logging mechanism for better production code practices.

print(f"Adding batch {i} to {i+batch_size} of {num_splits} with {len(splits_batch)} documents")

[Copilot]

Copilot reviewed 31 out of 46 changed files in this pull request and generated no comments.

Files not reviewed (15)

• .vscode/settings.json: Language not supported
• infra/deploy.sh: Language not supported
• src/api_input/requirements.txt: Language not supported
• src/api_input/start.sh: Language not supported
• src/autopodcaster_model/requirements.txt: Language not supported
• src/indexer/start.sh: Language not supported
• src/indexer/test.http: Language not supported
• src/pdf_indexer/requirements.txt: Language not supported
• src/podcast_generator/requirements.txt: Language not supported
• src/ui/AutoPodcaster/.vscode/launch.json: Language not supported
• src/indexer/README.md: Evaluated as low risk
• src/indexer/indexer.py: Evaluated as low risk
• src/api_input/api_input.py: Evaluated as low risk
• src/autopodcaster_model/autopodcaster_model/init.py: Evaluated as low risk
• src/autopodcaster_model/autopodcaster_model/input.py: Evaluated as low risk

Comments suppressed due to low confidence (1)

src/podcast_generator/podcast_generator.py:20

• The word 'imrpoved' should be corrected to 'improved'.

ONLY return the imrpoved modified text!!