Voice & Transcription (1.0.0)

Get clinical documentation instructions. This endpoint provides instruction sets for various clinical documentation tasks.

Args: instruction_type: Type of instruction (clinical_notes, safety_plan, etc.) user_id: Optional user ID for personalized instructions

Returns: Instructions for the specified documentation task

Accepts an audio file together with session metadata, uploads it to Azure Blob Storage under a structured path of the form ///, persists a TherapistSessionRecords entry to the database, and asynchronously triggers transcript generation via a Celery task. If a blob with the same name already exists in the container a numeric suffix (-1, -2, …) is appended to guarantee uniqueness. The is_last_blob flag signals the pipeline that all audio chunks for the session have been received and full transcription can begin.

Looks up a blob by its storage path (file_url) and streams it back to the caller as a chunked HTTP response with the original content type preserved. The Content-Disposition header is set to 'inline' so browsers render the file in-place rather than downloading it. Content-Length is also returned so clients can display accurate progress. Intended for use when rendering previously uploaded documents (e.g. intake forms, clinical notes) inside the application.

Accepts one or more document files alongside user metadata, uploads each file to Azure Blob Storage under the path ////, and immediately runs Azure Document Intelligence to extract the full text content. Image files (.png, .jpg, .jpeg, .tiff, .bmp) are processed with the 'prebuilt-read' OCR model; PDF files are processed with the 'prebuilt-layout' model for full structural analysis. The extracted text, original filename, and blob path are returned for every file in the batch. Unsupported file extensions raise a ValueError before upload.

Returns all non-deleted TherapistSessionTranscriptHistory records for the given client user ID and therapist-client relationship ID, ordered by most recent session first (descending create_timestamp). Each record contains an AES-encrypted, Base64-encoded transcript summary along with identifying metadata such as client/therapist user IDs and the associated session reference. Returns 204 when no history records exist.

Lists all audio blobs stored under the path /// in Azure Blob Storage, generates a short-lived SAS token for each blob, and synchronously transcribes every file using the AzureTranscriptionService. All resulting transcript strings are collected and returned in a single response. This endpoint is intended for manual or on-demand re-processing; automated transcription is normally triggered asynchronously via the Celery pipeline after the last blob is uploaded.

Returns all non-deleted TherapistSessionTranscript records that match the given client user ID and therapist-client relationship ID. Each record includes the full transcript and AI-generated summary, both AES-encrypted and Base64-encoded in the response. Returns 204 when no transcripts exist for the specified pair.