# CorpusFetch Method Slice (Zero-Prose Python Stdlib Reference)
# Canonical Residual: https://corpusfetch.com/c/what-embedding-metadata-invariants-must-survive-schema-evolution-in-vector-indices.md
# Remnant Warrant:    https://corpusfetch.com/c/what-embedding-metadata-invariants-must-survive-schema-evolution-in-vector-indices.warrant.json
# Markdown SHA-256:   38abab2c9abd0facaaf0afd98f68c32b8bb6effed93471ae16dbe9b34dae75c5
# Code AST SHA-256:   413e373313dbda362cd9003104c80a73edff01dc84677b335d478a42ca265669
# Verified Classes:   MetadataVerdict, VectorMetadata, SchemaContext, MetadataPolicy, MetadataManifest, EmbeddingMetadataGuard
# Defensive Guards:   0 fail-closed invariants (zero external deps)

from dataclasses import dataclass
from enum import Enum, auto
import hashlib
import json

class MetadataVerdict(Enum):
    SINGLE_VERSION = auto()
    CROSS_VERSION_TRANSLATED = auto()
    NO_TRANSLATION_MATRIX = auto()
    UNVERIFIED_VERSION = auto()

@dataclass(frozen=True)
class VectorMetadata:
    vector_id: str
    schema_version: str
    schema_signature: str | None

@dataclass(frozen=True)
class SchemaContext:
    vectors: tuple[VectorMetadata, ...]
    current_query_schema_version: str
    field_mapping_available: bool

@dataclass(frozen=True)
class MetadataPolicy:
    require_translation_for_cross_version: bool

@dataclass(frozen=True)
class MetadataManifest:
    verdict: MetadataVerdict
    directive: str
    manifest_digest: str

class EmbeddingMetadataGuard:
    @staticmethod
    def _hash(data: str) -> str:
        return hashlib.sha256(data.encode("utf-8")).hexdigest()

    @classmethod
    def evaluate_metadata(
        cls, ctx: SchemaContext, policy: MetadataPolicy
    ) -> MetadataManifest:
        
        # Gate 1: Signature presence
        for v in ctx.vectors:
            if not v.schema_signature:
                return MetadataManifest(
                    verdict=MetadataVerdict.UNVERIFIED_VERSION,
                    directive=f"HALT: Vector {v.vector_id} lacks schema version signature.",
                    manifest_digest=cls._hash(f"NO_SIG:{v.vector_id}")
                )

        # Gate 2: Version diversity check
        versions = {v.schema_version for v in ctx.vectors}

        # Gate 3: Single version
        if len(versions) == 1 and ctx.current_query_schema_version in versions:
            digest = cls._hash(json.dumps({
                "version": ctx.current_query_schema_version,
                "vectors": str(len(ctx.vectors))
            }, sort_keys=True))
            return MetadataManifest(
                verdict=MetadataVerdict.SINGLE_VERSION,
                directive=f"PROCEED: All vectors are schema version {ctx.current_query_schema_version}.",
                manifest_digest=digest
            )

        # Gate 4: Cross-version translation
        if policy.require_translation_for_cross_version and not ctx.field_mapping_available:
            return MetadataManifest(
                verdict=MetadataVerdict.NO_TRANSLATION_MATRIX,
                directive=f"HALT: Vectors span {len(versions)} schema versions, no translation matrix.",
                manifest_digest=cls._hash(f"NO_MATRIX:{len(versions)}")
            )

        # Gate 5: Translated
        digest = cls._hash(json.dumps({
            "versions": tuple(sorted(versions)),
            "query_version": ctx.current_query_schema_version,
            "vectors": str(len(ctx.vectors)),
            "translation": str(ctx.field_mapping_available)
        }, sort_keys=True))
        
        return MetadataManifest(
            verdict=MetadataVerdict.CROSS_VERSION_TRANSLATED,
            directive=f"PROCEED: Translated {len(versions)} schema versions to {ctx.current_query_schema_version}.",
            manifest_digest=digest
        )