From 9a5cd8b86e41a3eae5267168a067fba478faa2f7 Mon Sep 17 00:00:00 2001
From: Cyber MacGeddon <cybermaggedon@gmail.com>
Date: Wed, 16 Oct 2024 23:30:51 +0100
Subject: [PATCH] Building document metadata

---
 trustgraph-base/trustgraph/schema/metadata.py | 12 +++--
 trustgraph-cli/scripts/tg-load-pdf            | 54 ++++++++++++++++++-
 2 files changed, 62 insertions(+), 4 deletions(-)

diff --git a/trustgraph-base/trustgraph/schema/metadata.py b/trustgraph-base/trustgraph/schema/metadata.py
index c7dbbae6..aac38f46 100644
--- a/trustgraph-base/trustgraph/schema/metadata.py
+++ b/trustgraph-base/trustgraph/schema/metadata.py
@@ -1,10 +1,16 @@
 
-from pulsar.schema import Record, String
+from pulsar.schema import Record, String, Array
+from . types import Triple
 
 class Metadata(Record):
-    source = String()
+
+    # Source identifier
     id = String()
-    title = String()
+
+    # Subgraph
+    source = Array(Triple())
+
+    # Collection management
     user = String()
     collection = String()
 
diff --git a/trustgraph-cli/scripts/tg-load-pdf b/trustgraph-cli/scripts/tg-load-pdf
index 460a2f06..e0904609 100755
--- a/trustgraph-cli/scripts/tg-load-pdf
+++ b/trustgraph-cli/scripts/tg-load-pdf
@@ -55,7 +55,11 @@ class Loader:
             path = file
             data = open(path, "rb").read()
 
-            id = hashlib.sha256(path.encode("utf-8")).hexdigest()[0:8]
+            # Create a SHA256 hash from the data
+            id = hashlib.sha256(data).hexdigest()
+
+            # Convert into a UUID, 64-byte hash becomes 32-byte UUID
+            id = str(uuid.UUID(id[::2]))
 
             r = Document(
                 metadata=Metadata(
@@ -112,6 +116,54 @@ def main():
         help=f'Collection ID (default: {default_collection})'
     )
 
+    parser.add_argument(
+        '--name', help=f'Document name'
+    )
+
+    parser.add_argument(
+        '--description', help=f'Document description'
+    )
+
+    parser.add_argument(
+        '--copyright-notice', help=f'Copyright notice'
+    )
+
+    parser.add_argument(
+        '--copyright-holder', help=f'Copyright holder'
+    )
+
+    parser.add_argument(
+        '--copyright-year', help=f'Copyright year'
+    )
+
+    parser.add_argument(
+        '--licence', help=f'Copyright licence'
+    )
+
+    parser.add_argument(
+        '--publication-organization', help=f'Publication organization'
+    )
+
+    parser.add_argument(
+        '--publication-description', help=f'Publication description'
+    )
+
+    parser.add_argument(
+        '--publication-date', help=f'Publication date'
+    )
+
+    parser.add_argument(
+        '--url', help=f'Document URL'
+    )
+
+    parser.add_argument(
+        '--keyword', help=f'Keyword'
+    )
+
+    parser.add_argument(
+        '--title', help=f'Document title'
+    )
+
     parser.add_argument(
         '-l', '--log-level',
         type=LogLevel,