clio/clio-api/importer.py

"""
Import functionality for recordings and transcriptions.

Supports idempotent import of:
- JSON transcription files (from transcribe.py output)
- Audio files from various sources
"""

import json
import os
import re
from datetime import datetime
from pathlib import Path
from typing import Optional

import database


def parse_timestamp_from_filename(filename: str) -> Optional[datetime]:
    """
    Extract timestamp from filename like 'recording_2026-04-26_19-52-47.json'.
    Returns None if pattern doesn't match.
    """
    pattern = r'recording_(\d{4})-(\d{2})-(\d{2})_(\d{2})-(\d{2})-(\d{2})'
    match = re.search(pattern, filename)
    if match:
        year, month, day, hour, minute, second = map(int, match.groups())
        return datetime(year, month, day, hour, minute, second)
    return None


def import_transcription_json(
    json_path: Path,
    source_name: str = "phone",
    audio_base_path: Optional[Path] = None,
) -> dict:
    """
    Import a single transcription JSON file.

    Returns dict with:
        - status: 'created', 'exists', or 'error'
        - recording_id: ID if created or exists
        - message: description
    """
    try:
        with open(json_path) as f:
            data = json.load(f)
    except (json.JSONDecodeError, IOError) as e:
        return {"status": "error", "message": f"Failed to read JSON: {e}"}

    source_file = data.get("source_file", str(json_path))

    # Check if already imported (idempotent)
    existing = database.get_recording_by_source_file(source_file)
    if existing:
        return {
            "status": "exists",
            "recording_id": existing["id"],
            "message": f"Recording already imported: {existing['filename']}",
        }

    # Parse filename and timestamp
    filename = Path(source_file).name
    recorded_at = parse_timestamp_from_filename(filename)

    # Determine audio path
    audio_path = None
    if audio_base_path:
        # Try to find audio file relative to base path
        audio_file = Path(source_file)
        if audio_file.exists():
            audio_path = str(audio_file)
        else:
            # Try in audio_base_path
            potential = audio_base_path / audio_file.name
            if potential.exists():
                audio_path = str(potential)

    # Create recording entry
    recording_id = database.create_recording(
        source_file=source_file,
        filename=filename,
        source_name=source_name,
        audio_path=audio_path,
        duration=data.get("duration"),
        language=data.get("language"),
        language_probability=data.get("language_probability"),
        recorded_at=recorded_at,
        transcribed_at=datetime.now(),
    )

    # Import segments
    segments = data.get("segments", [])
    segment_count = database.create_segments(recording_id, segments)

    return {
        "status": "created",
        "recording_id": recording_id,
        "message": f"Imported {filename} with {segment_count} segments",
        "segments": segment_count,
    }


def import_transcription_directory(
    directory: Path,
    source_name: str = "phone",
    audio_base_path: Optional[Path] = None,
) -> dict:
    """
    Import all JSON transcription files from a directory.

    Returns summary with counts of created, existing, and errors.
    """
    results = {
        "created": 0,
        "exists": 0,
        "errors": 0,
        "details": [],
    }

    json_files = sorted(directory.glob("*.json"))

    for json_path in json_files:
        result = import_transcription_json(json_path, source_name, audio_base_path)
        results["details"].append({
            "file": json_path.name,
            **result,
        })

        if result["status"] == "created":
            results["created"] += 1
        elif result["status"] == "exists":
            results["exists"] += 1
        else:
            results["errors"] += 1

    return results


def import_audio_file(
    audio_path: Path,
    source_name: str,
    move_to: Optional[Path] = None,
) -> dict:
    """
    Import a single audio file (without transcription).

    If move_to is provided, the file will be moved there.
    Otherwise, just creates a database entry pointing to current location.
    """
    if not audio_path.exists():
        return {"status": "error", "message": f"File not found: {audio_path}"}

    source_file = str(audio_path.resolve())

    # Check if already imported
    existing = database.get_recording_by_source_file(source_file)
    if existing:
        return {
            "status": "exists",
            "recording_id": existing["id"],
            "message": f"Audio already imported: {existing['filename']}",
        }

    filename = audio_path.name
    recorded_at = parse_timestamp_from_filename(filename)

    # Get file size
    file_size = audio_path.stat().st_size

    # Move file if requested
    final_path = str(audio_path.resolve())
    if move_to:
        move_to.mkdir(parents=True, exist_ok=True)
        dest = move_to / filename
        audio_path.rename(dest)
        final_path = str(dest.resolve())

    recording_id = database.create_recording(
        source_file=source_file,
        filename=filename,
        source_name=source_name,
        audio_path=final_path,
        recorded_at=recorded_at,
    )

    # Update file size
    database.update_recording(recording_id, file_size_bytes=file_size)

    return {
        "status": "created",
        "recording_id": recording_id,
        "message": f"Imported audio: {filename}",
    }


def import_audio_directory(
    directory: Path,
    source_name: str,
    extensions: tuple[str, ...] = (".ogg", ".mp3", ".wav", ".m4a", ".flac", ".webm"),
    move_to: Optional[Path] = None,
) -> dict:
    """
    Import all audio files from a directory.
    """
    results = {
        "created": 0,
        "exists": 0,
        "errors": 0,
        "details": [],
    }

    # Recursively find all audio files
    audio_files = sorted(
        p for ext in extensions
        for p in directory.rglob(f"*{ext}")
    )

    for audio_path in audio_files:
        result = import_audio_file(audio_path, source_name, move_to)
        results["details"].append({
            "file": audio_path.name,
            **result,
        })

        if result["status"] == "created":
            results["created"] += 1
        elif result["status"] == "exists":
            results["exists"] += 1
        else:
            results["errors"] += 1

    return results


def link_audio_to_transcriptions(audio_dir: Path) -> dict:
    """
    Link audio files to existing transcription records that are missing audio_path.

    Matches by filename pattern (e.g., recording_2026-04-26_19-52-47.ogg).
    """
    updated = 0
    not_found = 0

    with database.get_db() as conn:
        # Get all recordings without audio_path
        cur = conn.execute(
            "SELECT id, filename, source_file FROM recordings WHERE audio_path IS NULL"
        )
        recordings = cur.fetchall()

        for rec in recordings:
            # Try to find matching audio file
            base_name = Path(rec["filename"]).stem
            for ext in (".ogg", ".mp3", ".wav", ".m4a", ".flac", ".webm"):
                potential = audio_dir / f"{base_name}{ext}"
                if potential.exists():
                    conn.execute(
                        "UPDATE recordings SET audio_path = ? WHERE id = ?",
                        (str(potential.resolve()), rec["id"])
                    )
                    updated += 1
                    break
            else:
                not_found += 1

    return {
        "updated": updated,
        "not_found": not_found,
        "message": f"Linked {updated} recordings to audio files, {not_found} still missing",
    }