ThothII/harness/tht/cli/vector_cmd.py

from pathlib import Path

import typer

from tht.cli._guards import has_vector_write_rest, require_server_profile, require_vector_write_allowed
from tht.cli.config_cmd import CONFIG_OPT
from tht.cli.schema_cmd import _load_config_or_exit, annotations_path, physical_path

vector_app = typer.Typer(help="Indice semantico pgvector (derivato, rigenerabile)")


def make_embedder(embeddings_cfg):
    """Factory del client embeddings (monkeypatchabile nei test)."""
    from tht.vectorstore.embeddings import OllamaEmbeddings

    return OllamaEmbeddings(embeddings_cfg)


def require_vector_cfg(cfg):
    missing = []
    if cfg.embeddings is None:
        missing.append("embeddings")
    if cfg.vector_db is None and not has_vector_write_rest(cfg):
        missing.append("vector_db o vector_write_rest")
    if missing:
        typer.secho(
            f"ERRORE: sezioni mancanti nel workspace yaml: {', '.join(missing)}.",
            fg=typer.colors.RED, err=True,
        )
        raise typer.Exit(code=1)


def require_direct_vector_cfg(cfg):
    missing = [k for k in ("vector_db", "embeddings") if getattr(cfg, k) is None]
    if missing:
        typer.secho(
            f"ERRORE: sezioni mancanti nel workspace yaml: {', '.join(missing)}.",
            fg=typer.colors.RED, err=True,
        )
        raise typer.Exit(code=1)


def open_store(cfg, table: str):
    """Writer table-scoped per il LOADING.

    Sul server preferisce la connessione diretta. In profilo workstation usa `vector_write_rest`
    se configurato, con upsert remoto non distruttivo.
    """
    if has_vector_write_rest(cfg) and (cfg.profile == "workstation" or cfg.vector_db is None):
        from tht.vectorstore.rest_client import VectorRestClient
        from tht.vectorstore.rest_writer import RestVectorWriter

        return RestVectorWriter(VectorRestClient(cfg.vector_write_rest), table=table)

    from tht.db.connection import make_engine
    from tht.vectorstore.store import VectorStore

    engine = make_engine(cfg.vector_db)
    return VectorStore(
        engine, schema=cfg.vector_db.db_schema, table=table, dim=cfg.embeddings.dim
    )


def open_searcher(cfg):
    """Searcher per la LETTURA (similarity search): via REST se `vector_rest` è configurato,
    altrimenti connessione diretta (dev/test)."""
    if cfg.vector_rest is not None:
        from tht.vectorstore.reader import RestSearcher
        from tht.vectorstore.rest_client import VectorRestClient

        return RestSearcher(VectorRestClient(cfg.vector_rest))
    from tht.db.connection import make_engine
    from tht.vectorstore.reader import DirectSearcher

    return DirectSearcher(
        make_engine(cfg.vector_db), schema=cfg.vector_db.db_schema, dim=cfg.embeddings.dim
    )


def _print_stats(stats) -> None:
    typer.secho(
        f"OK: {stats.added} nuovi, {stats.updated} aggiornati, "
        f"{stats.deleted} rimossi, {stats.unchanged} invariati",
        fg=typer.colors.GREEN,
    )


@vector_app.command("init")
def init_cmd(
    config: Path = CONFIG_OPT,
    skip_ollama_check: bool = typer.Option(
        False, "--skip-ollama-check", help="Non verificare la raggiungibilita' di Ollama."
    ),
) -> None:
    """Crea schema e tabella pgvector (idempotente) e verifica le connessioni."""
    from sqlalchemy.exc import OperationalError

    from tht.vectorstore.embeddings import EmbeddingsError

    from tht.vectorstore.reader import ALL_TABLES

    cfg = _load_config_or_exit(config)
    require_server_profile(cfg, "vector init")
    require_direct_vector_cfg(cfg)
    try:
        for table in ALL_TABLES:
            open_store(cfg, table).init_schema()
    except OperationalError as e:
        typer.secho(f"ERRORE connessione pgvector: {e.orig}", fg=typer.colors.RED, err=True)
        raise typer.Exit(code=1)
    if not skip_ollama_check:
        try:
            make_embedder(cfg.embeddings).embed_query("ping")
        except EmbeddingsError as e:
            typer.secho(f"ERRORE: {e}", fg=typer.colors.RED, err=True)
            raise typer.Exit(code=1)
    typer.secho(
        f"OK: schema {cfg.vector_db.db_schema} pronto (tabelle: {', '.join(ALL_TABLES)}) su "
        f"{cfg.vector_db.host}:{cfg.vector_db.port}", fg=typer.colors.GREEN,
    )


@vector_app.command("index-schema")
def index_schema_cmd(config: Path = CONFIG_OPT) -> None:
    """Embedda e sincronizza i record schema (tabelle e colonne) da mschema."""
    from tht.mschema.models import Annotations, PhysicalSchema
    from tht.vectorstore.records import schema_records

    cfg = _load_config_or_exit(config)
    require_vector_write_allowed(cfg, "vector index-schema")
    require_vector_cfg(cfg)
    phys_file = physical_path(cfg)
    if not phys_file.exists():
        typer.secho(
            f"ERRORE: {phys_file} non trovato. Esegui prima `tht schema introspect`.",
            fg=typer.colors.RED, err=True,
        )
        raise typer.Exit(code=1)
    physical = PhysicalSchema.from_yaml(phys_file)
    annotations = Annotations.from_yaml(annotations_path(cfg))
    records = schema_records(physical, annotations)
    store = open_store(cfg, "schema_records")
    stats = store.sync(
        records, make_embedder(cfg.embeddings), kinds={"schema_table", "schema_column"}
    )
    _print_stats(stats)