AI Copilot API Endpoints Documentation

[Grinzypino]

AI Copilot - Flask App API Endpoints Documentation

This documentation provides an overview of the routes in a Flask application that interacts with OpenAI's GPT models and Azure's translation services. The application handles generating questions, reframing questions, checking answers, and translating content.

Table of Contents

• Setup
• Routes
  • POST /generate_draft
  • POST /reframe_q
  • POST /check_ans
  • POST /generate_q
• Helper Functions

Setup

To set up the application, ensure you have the necessary environment variables and dependencies.

1. Install the required packages:

   Copy code

   pip install flask openai python-dotenv requests

2. Set the necessary environment variables in a .env file:

   makefile

   Copy code

   AZURE_SUBSCRIPTION_KEY=your_azure_subscription_key
OPENAI_API_KEY=your_openai_key

Routes

POST /generate_q

Generates initial questions based on the provided schema path.

• Request: JSON
  • schema_path: Path to the JSON schema file.
• Response: JSON
  • q_dict: Dictionary of generated questions.
  • current_json: Null or empty JSON object based on the schema.
  • expected_dict_ans: Dictionary of expected answers.

Example:

Request JSON:

{ "schema_path": "schemaNexamples.json" }

Response JSON:
{ "current_json": { "additional_documents": [], "affidavit_submitted": null, "annulment_exists": null, "tehsil": null }, "expected_dict_ans": { "additional_documents": { "example": [ "extract Khatauni", "affidavit", "original sale deed" ], "items": { "type": "string" }, "type": "array" }, "affidavit_submitted": { "example": [ true, false ], "type": "boolean" }, "annulment_exists": { "example": [ true, false ], "type": "boolean" }, "tehsil": { "example": [ "Sadar Tehsil", "Sadar", "Tehsil Sadar" ], "type": "string" } }, "q_dict": { "additional_documents": "What are the additional documents submitted?", "affidavit_submitted": "Was an affidavit submitted?", "annulment_exists": "Does annulment of the case exist?", "annulment_order_date": "On what date was the annulment ordered?", } }

POST /check_ans

Checks and updates the current JSON object with the provided answers by user.

• Request: JSON
  • q_dict: Dictionary of questions.
  • expected_dict_ans: Dictionary of expected answers.
  • current_json: The current JSON object.
  • user_ans: The user's answers in Hindi.
• Response: JSON
  • q_dict_unans: Dictionary of unanswered questions.
  • current_json: Updated current JSON object.
  • expected_dict_ans: Dictionary of expected answers for the unanswered questions.

Example:
Request JSON:
{ "current_json": { "additional_documents": [], "affidavit_submitted": null, "annulment_exists": null, "tehsil": null }, "expected_dict_ans": { "additional_documents": { "example": [ "extract Khatauni", "affidavit", "original sale deed" ], "items": { "type": "string" }, "type": "array" }, "affidavit_submitted": { "example": [ true, false ], "type": "boolean" }, "annulment_exists": { "example": [ true, false ], "type": "boolean" }, "tehsil": { "example": [ "Sadar Tehsil", "Sadar", "Tehsil Sadar" ], "type": "string" } }, "q_dict": { "additional_documents": "What are the additional documents submitted?", "affidavit_submitted": "Was an affidavit submitted?", "annulment_exists": "Does annulment of the case exist?", "tehsil": "Which Tehsil does the land belong to?" }, "user_ans": { "tehsil": "सदर तहसील", "annulment_exists": true, "affidavit_submitted": true, "additional_documents": [], } }

Response JSON:
{ "current_json": { "additional_documents": [], "affidavit_submitted": true, "annulment_exists": true, "tehsil": "Sadar Tehsil" }, "expected_dict_ans": { "additional_documents": { "example": [ "extract Khatauni", "affidavit", "original sale deed" ], "items": { "type": "string" }, "type": "array" } } }, "q_dict_unans": { "additional_documents": "What are the additional documents submitted?", } }

POST /reframe_q

Reframes the unanswered questions based on the expected answers.

• Request: JSON
  • Q_dict_unans: Dictionary of unanswered questions.
  • expected_dict_ans: Dictionary of expected answers for the unanswered questions.
• Response: JSON
  • q_dict: Dictionary of reframed questions.

Example:

Request JSON:
{ "q_dict_unans": { "additional_documents": "What are the additional documents submitted?" }, "expected_dict_ans": { "additional_documents": { "example": [ "extract Khatauni", "affidavit", "original sale deed" ], "items": { "type": "string" }, "type": "array" } } }

Response JSON:
{ "q_dict": { "additional_documents": "What are some examples of additional documents that may be required? (e.g. extract Khatauni, affidavit, original sale deed)" } }

POST /generate_draft

Generates a draft response based on the provided JSON schema.

• Request: JSON
  • current_json: The current JSON object that needs to be processed.
• Response: JSON
  • eng_draft_path: Path to the English draft text file.
  • hin_draft_path: Path to the translated Hindi draft text file.

Example:

Request JSON:
{ "current_json": { "additional_documents": [ "extract khatauuni", "affidavit", "Original Sale Deed" ], "affidavit_submitted": true, "annulment_exists": true, "tehsil": "Sadar Tehsil" } }

Response JSON:
{ "eng_draft_path": "Draft.txt", "hin_draft_path": "DraftTranslated.txt" }

Helper Functions

• load_json(file_path): Loads a JSON file from the specified file path.
• generate_questions(schema_path): Generates questions based on a JSON schema.
• fill_JSON_schema(user_ans_str, already_filled_str): Fills values in the current JSON schema based on user input.
• reframe_Q(Q_dict_earlier, expected_dict_ans): Reframes questions based on expected answers.
• convert_q(q_str): Converts questions from English to Hindi.
• convert_ans(hin_ans): Converts answers from Hindi to English.
• print_q(q_str): Prints questions.
• print_ans(ans_str): Prints answers.
• check_dict(d, q_dict, expected): Checks the dictionary for unanswered questions.
• create_null_json(input_json_path): Creates a null or empty JSON object based on the input schema.

These helper functions are used internally by the routes to process and manage the JSON data.