API Reference

API Reference

This section details the API endpoints provided by the Chat-Service. These are primarily defined in the src/routes/ directory.

Endpoints

Debug Logs

Endpoints defined in src/routes/debug.py.

• GET /api/debug/logs
  • Description: Returns in-memory logs for progress and failures. Useful for debugging the service's current state.
  • Request Body: None.
  • Response:

    {
      "progressLog": [ /* list of progress entries */ ],
      "failureLog": [ /* list of failure entries */ ]
    }

Progress

Endpoints defined in src/routes/progress.py. These are used for tracking task completion, potentially for user-facing objectives or internal processes.

• POST /api/progress/initializeTasks

  • Description: Initializes a list of tasks, including their subtasks and steps. This can be used to set up a new set of trackable items.
  • Request Body:

    {
      "items": [
        {
          "taskName": "string",
          "status": "pending", // or "started", "complete"
          "subtaskProgress": [
            {
              "subtaskName": "string",
              "status": "pending" // or "started", "complete"
            }
          ],
          "startedAt": "datetime string (optional)",
          "completedAt": "datetime string (optional)"
        }
      ]
    }

  • Response:

    {
      "message": "Tasks initialized",
      "data": { /* the taskHierarchy sent in the request */ }
    }

• POST /api/progress/updateTask

  • Description: Handles updates to task progress (e.g., marking as started or complete).
  • Request Body:

    {
      "taskName": "string",
      "status": "started", // or "pending", "complete"
      "subtaskProgress": [ /* same structure as in initializeTasks */ ],
      "startedAt": "datetime string (optional)",
      "completedAt": "datetime string (optional)"
    }

  • Response:

    {
      "message": "Progress updated" / "Progress received" / "Task completed",
      "data": { /* updated or new task entry */ }
    }

  • Notes: If a task is "started" or "pending" and already exists as incomplete, it's updated. If new, it's added. If "complete", an existing active task is marked as completed.

• GET /api/progress

  • Description: Returns the entire in-memory log of progress entries.
  • Request Body: None.
  • Response:

    [ /* list of progress data entries */ ]

Upload

Endpoint defined in src/routes/upload.py. Used for uploading documents for context, used for the RAG Service.

• POST /upload/
  • Description: Receives a document file, processes it (e.g., for embedding), and stores it.
  • Request Parameters:
    • file: The document file to upload (e.g., .txt, .md). Sent as multipart/form-data.
    • category (optional query parameter): The category of the document. Defaults to "General Information".
      • Possible values include: "Behavioral Context", "General Information", "Conversational Assets", "Training Data", "Miscellaneous", "FishFeeding", "Laboratory", "FishFactory", "FishWelfare", "FishMaintenance", "Ocean".
  • Response (Success):

    {
      "message": "File uploaded and processed successfully",
      "category": "string" // The category used
    }

  • Response (Error):

    {
      "detail": "Error message"
    }

    (HTTP 500 for processing failures, or other relevant HTTP error codes)

Chat / Query (Placeholder)

This endpoint's definition and location are yet to be fully confirmed from the codebase. It's expected to handle interactions with the LLM.

• POST /chat (Example Path)
  • Description: Sends text to the LLM for a response, potentially using RAG.
  • Request Body (Example):

    {
      "question": "User's query.",
      "user_information": ["unique_user_identifier"],
      "progress": [ /* list of ProgressData objects - definition TBD */ ],
      "user_actions": [ /* list of strings describing user actions */ ],
      "base_prompt": "System-level instructions for the LLM.",
      "context": "Relevant contextual information as a string."
    }

  • Response (Example):

    {
      "response_text": "LLM generated response.",
      "source_documents": [ /* If RAG is used */ ]
    }