Legal corpus: seed U.S. Code Title 52

2026-07-04 19:49:09 -04:00
parent fe8f82ae30
commit ad7607856d
182 changed files with 22116 additions and 16 deletions
@@ -0,0 +1,35 @@
+{
+  "$schema": "http://json-schema.org/draft-07/schema#",
+  "title": "LegalManifest",
+  "description": "Audit manifest for a deterministic legal corpus import.",
+  "type": "object",
+  "required": [
+    "type",
+    "corpus",
+    "jurisdiction",
+    "release_point",
+    "source",
+    "source_url",
+    "source_file_sha256",
+    "section_count",
+    "files"
+  ],
+  "properties": {
+    "type": { "type": "string", "enum": ["LegalManifest"] },
+    "corpus": { "type": "string", "enum": ["united_states_code"] },
+    "jurisdiction": { "type": "string" },
+    "title_number": { "type": "integer" },
+    "title_name": { "type": "string" },
+    "release_point": { "type": "string" },
+    "release_date": { "type": "string" },
+    "retrieved_at": { "type": "string" },
+    "source": { "type": "string", "enum": ["official"] },
+    "source_url": { "type": "string" },
+    "source_file": { "type": "string" },
+    "source_file_sha256": { "type": "string" },
+    "xml_file": { "type": "string" },
+    "xml_file_sha256": { "type": "string" },
+    "section_count": { "type": "integer" },
+    "files": { "type": "array" }
+  }
+}
@@ -0,0 +1,51 @@
+{
+  "$schema": "http://json-schema.org/draft-07/schema#",
+  "title": "LegalText",
+  "description": "A deterministic, source-backed legal text node such as a codified law section.",
+  "type": "object",
+  "required": [
+    "type",
+    "title",
+    "jurisdiction",
+    "corpus",
+    "kind",
+    "citation",
+    "status",
+    "source",
+    "source_url",
+    "source_identifier",
+    "source_hash",
+    "raw_snapshot_hash",
+    "text_hash",
+    "retrieved_at",
+    "confidence",
+    "tags"
+  ],
+  "properties": {
+    "type": { "type": "string", "enum": ["LegalText"] },
+    "title": { "type": "string" },
+    "description": { "type": "string" },
+    "jurisdiction": { "type": "string" },
+    "corpus": { "type": "string", "enum": ["united_states_code"] },
+    "kind": { "type": "string", "enum": ["code_section"] },
+    "title_number": { "type": "integer" },
+    "title_name": { "type": "string" },
+    "chapter_number": { "type": "string" },
+    "chapter_name": { "type": "string" },
+    "section": { "type": "string" },
+    "citation": { "type": "string" },
+    "status": { "type": "string", "enum": ["current", "repealed", "superseded", "reserved"] },
+    "release_point": { "type": "string" },
+    "release_date": { "type": "string" },
+    "source": { "type": "string", "enum": ["official"] },
+    "source_url": { "type": "string" },
+    "source_identifier": { "type": "string" },
+    "source_file": { "type": "string" },
+    "source_hash": { "type": "string" },
+    "raw_snapshot_hash": { "type": "string" },
+    "text_hash": { "type": "string" },
+    "retrieved_at": { "type": "string" },
+    "confidence": { "type": "string", "enum": ["official", "reported", "inferred", "unverified"] },
+    "tags": { "type": "array" }
+  }
+}