"""Reusable vectors keyed by the exact input and model fingerprint."""

import hashlib
import json
import sqlite3
from contextlib import closing
from pathlib import Path

import numpy as np


def cache_key(profile: dict, text: str) -> str:
    return hashlib.sha256(
        json.dumps(
            {"profile": profile["fingerprint"], "text": text}, ensure_ascii=False, sort_keys=True
        ).encode()
    ).hexdigest()


class EmbeddingCache:
    def __init__(self, path: Path, *, readonly: bool = False):
        self.path = Path(path)
        self.readonly = readonly
        if readonly:
            return
        self.path.parent.mkdir(parents=True, exist_ok=True)
        with closing(sqlite3.connect(self.path)) as conn, conn:
            conn.execute("""CREATE TABLE IF NOT EXISTS embeddings (
                key TEXT PRIMARY KEY, dimensions INTEGER NOT NULL, vector BLOB NOT NULL,
                vector_sha256 TEXT NOT NULL)""")

    def load(self, keys: list[str]) -> dict[str, np.ndarray]:
        found = {}
        target = self.path.resolve().as_uri() + "?mode=ro" if self.readonly else str(self.path)
        with closing(sqlite3.connect(target, uri=self.readonly)) as conn:
            for key in set(keys):
                row = conn.execute(
                    "SELECT dimensions,vector,vector_sha256 FROM embeddings WHERE key=?", (key,)
                ).fetchone()
                if not row or hashlib.sha256(row[1]).hexdigest() != row[2]:
                    continue
                vector = np.frombuffer(row[1], dtype=np.float32)
                if (
                    len(vector) == row[0]
                    and np.isfinite(vector).all()
                    and np.linalg.norm(vector) > 0
                ):
                    found[key] = vector.copy()
        return found

    def save(self, vectors: dict[str, np.ndarray]):
        if self.readonly:
            raise RuntimeError("Cannot write to a read-only embedding cache")
        with closing(sqlite3.connect(self.path)) as conn, conn:
            for key, vector in vectors.items():
                blob = np.asarray(vector, dtype=np.float32).tobytes()
                conn.execute(
                    "INSERT OR REPLACE INTO embeddings VALUES(?,?,?,?)",
                    (key, len(vector), blob, hashlib.sha256(blob).hexdigest()),
                )