Validator
Ce contenu n’est pas encore disponible dans votre langue.
Validate with the PAM SDK
Section titled “Validate with the PAM SDK”The official Python SDK is the recommended way to validate PAM files. It performs deep validation beyond schema compliance — verifying content hashes, cross-references, integrity blocks, and temporal consistency.
pip install 'portable-ai-memory[cli]'pam validate memory-store.jsonFor programmatic usage, bundle validation, and all available options, see the Validation Guide.
What the SDK validates
Section titled “What the SDK validates”| Check | What it verifies |
|---|---|
| Schema compliance | Required fields, correct types, valid enums |
| Content hashes | Each content_hash matches actual content per spec §6 normalization |
| Integrity block | total_memories count and aggregate checksum |
| Cross-references | Relations, conversation_ref, superseded_by, and derived_memories point to existing objects |
| Temporal ordering | created_at ≤ updated_at, valid_from ≤ valid_until |
| ID uniqueness | No duplicate memory, relation, or conversation IDs |
| Custom types | type='custom' requires custom_type and vice versa |
| Status consistency | superseded status ↔ superseded_by field |
| Conversation DAG | Message parent_id and children_ids consistency |
Without the SDK
Section titled “Without the SDK”If you can’t use Python, you can validate against the JSON Schemas directly with any JSON Schema Draft 2020-12 validator. See the Validation Guide for jsonschema and ajv examples. Manual validation only checks schema compliance.