feat: initial unicosys hypergraph knowledge model (34.7M params, 203K nodes, 15K edges)

README.md

@@ -0,0 +1,85 @@
+---
+license: mit
+tags:
+  - knowledge-graph
+  - hypergraph
+  - legal-evidence
+  - graph-neural-network
+  - unicosys
+language:
+  - en
+library_name: transformers
+pipeline_tag: graph-ml
+---
+
+# Unicosys Hypergraph Knowledge Model
+
+A trainable knowledge graph embedding model encoding the unified evidence
+hypergraph for Case 2025-137857.
+
+## Model Description
+
+This model encodes a **unified hypergraph** linking financial transactions,
+email communications, legal evidence, and entity relationships into a
+single trainable knowledge representation.
+
+### Architecture
+
+| Component | Details |
+|---|---|
+| Node Embedding | 128-dim structural + 256-dim text |
+| Hidden Dimension | 256 |
+| Text Encoder | 2-layer Transformer, 4 heads |
+| Graph Attention | 2-layer GAT, 4 heads |
+| Link Predictor | 2-layer MLP with margin ranking loss |
+| Total Parameters | **34,758,785** |
+
+### Knowledge Graph Statistics
+
+| Metric | Count |
+|---|---|
+| Total Nodes | 203,642 |
+| Total Edges | 14,675 |
+| Cross-Links | 4,116 |
+| Entities | 16 |
+| Emails | 199,541 |
+| Financial Documents | 2,290 |
+| Timeline Events | 1,220 |
+| LEX Schemes | 13 |
+| Legal Filings | 7 |
+
+### Subsystems
+
+| Subsystem | Nodes |
+|---|---|
+| Core (Entities) | 16 |
+| Fincosys (Financial) | 3,908 |
+| Comcosys (Communications) | 199,541 |
+| RevStream1 (Evidence) | 144 |
+| Ad-Res-J7 (Legal) | 33 |
+
+## Training
+
+The model can be fine-tuned on link prediction tasks:
+
+```python
+from model.unicosys_model import UnicosysHypergraphModel, UnicosysConfig
+
+model = UnicosysHypergraphModel.from_pretrained("hyperholmes/unicosys-hypergraph")
+# ... prepare training data ...
+# model.forward(node_ids, node_type_ids, subsystem_ids, edge_index, edge_type_ids,
+#               pos_edge_index=pos, neg_edge_index=neg, labels=labels)
+```
+
+## Files
+
+- `model.safetensors` — Model weights
+- `config.json` — Model configuration
+- `graph_data.safetensors` — Encoded graph tensors (nodes, edges)
+- `tokenizer.json` — Character-level tokenizer for node labels
+- `node_id_mapping.json` — Node ID string to integer index mapping
+- `model_summary.json` — Compact statistics summary
+
+## Source
+
+Generated by the [Unicosys](https://github.com/hyperholmes/unicosys) intelligence pipeline.

config.json

@@ -0,0 +1,55 @@
+{
+  "architectures": [
+    "UnicosysHypergraphModel"
+  ],
+  "case_number": "2025-137857",
+  "dtype": "float32",
+  "edge_type_vocab": {
+    "<pad>": 0,
+    "<unk>": 1,
+    "proves": 2,
+    "transaction_evidenced_by": 3
+  },
+  "gat_dropout": 0.1,
+  "gat_num_heads": 4,
+  "gat_num_layers": 2,
+  "hidden_dim": 256,
+  "margin": 1.0,
+  "max_nodes": 250000,
+  "model_type": "unicosys_hypergraph",
+  "negative_sample_ratio": 5,
+  "node_embed_dim": 128,
+  "node_type_vocab": {
+    "<pad>": 0,
+    "<unk>": 1,
+    "email": 2,
+    "entity": 3,
+    "entity_document": 4,
+    "fund_flow_analysis": 5,
+    "hypergraph_node": 6,
+    "legal_filing": 7,
+    "lex_scheme": 8,
+    "timeline_event": 9
+  },
+  "num_cross_links": 4116,
+  "num_edge_types": 4,
+  "num_entities": 16,
+  "num_evidence": 203642,
+  "num_node_types": 10,
+  "num_subsystems": 7,
+  "subsystem_vocab": {
+    "<pad>": 0,
+    "<unk>": 1,
+    "ad_res_j7": 2,
+    "comcosys": 3,
+    "core": 4,
+    "fincosys": 5,
+    "revstream1": 6
+  },
+  "text_embed_dim": 256,
+  "text_max_length": 128,
+  "text_num_heads": 4,
+  "text_num_layers": 2,
+  "text_vocab_size": 219,
+  "transformers_version": "5.3.0"
+}

graph_data.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33faa0851dda60392855cc055a950f565f22e40f55425c6f24460ae8cdb86f5a
+size 5264784

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e927a8a37ee0bcdc708ebfb937dace604b9a8d325516a19757173806cb25f63
+size 139041988

model_summary.json

@@ -0,0 +1,36 @@
+{
+  "case_number": "2025-137857",
+  "total_nodes": 203642,
+  "total_edges": 15707,
+  "total_cross_links": 4116,
+  "node_types": {
+    "entity": 16,
+    "entity_document": 2290,
+    "timeline_event": 1220,
+    "hypergraph_node": 551,
+    "fund_flow_analysis": 4,
+    "email": 199541,
+    "lex_scheme": 13,
+    "legal_filing": 7
+  },
+  "edge_types": {
+    "proves": 14674,
+    "transaction_evidenced_by": 1
+  },
+  "subsystems": {
+    "core": 16,
+    "fincosys": 3908,
+    "comcosys": 199541,
+    "revstream1": 144,
+    "ad_res_j7": 33
+  },
+  "model_params": 34758785,
+  "model_architecture": {
+    "node_embed_dim": 128,
+    "text_embed_dim": 256,
+    "hidden_dim": 256,
+    "gat_layers": 2,
+    "gat_heads": 4,
+    "text_encoder_layers": 2
+  }
+}

tokenizer.json

@@ -0,0 +1,122 @@
+{
+  "vocab_size": 32000,
+  "max_length": 128,
+  "char_to_id": {
+    "p": 4,
+    "e": 5,
+    "t": 6,
+    "r": 7,
+    " ": 8,
+    "a": 9,
+    "n": 10,
+    "d": 11,
+    "w": 12,
+    "f": 13,
+    "u": 14,
+    "c": 15,
+    "i": 16,
+    "j": 17,
+    "q": 18,
+    "l": 19,
+    "y": 20,
+    "o": 21,
+    "b": 22,
+    "s": 23,
+    "k": 24,
+    "v": 25,
+    "m": 26,
+    "h": 27,
+    "g": 28,
+    "(": 29,
+    ")": 30,
+    "\u00e9": 31,
+    "z": 32,
+    ":": 33,
+    "4": 34,
+    "8": 35,
+    "3": 36,
+    "1": 37,
+    "0": 38,
+    "7": 39,
+    "6": 40,
+    "9": 41,
+    "2": 42,
+    "5": 43,
+    "x": 44,
+    ",": 45,
+    "-": 46,
+    "/": 47,
+    "&": 48,
+    ".": 49,
+    "+": 50,
+    "%": 51,
+    "#": 52,
+    "'": 53,
+    "@": 54,
+    "_": 55,
+    "*": 56,
+    "|": 57,
+    "[": 58,
+    "]": 59,
+    "!": 60,
+    "\u2122": 61,
+    "=": 62,
+    "\u00f1": 63,
+    "\u263a": 64,
+    "\u00a3": 65,
+    "\u2013": 66,
+    "\ud83d\udd17": 67,
+    "\ud83d\udc9c": 68,
+    "\u00a0": 69,
+    "\u26f3": 70,
+    "\u26a1": 71,
+    "\u23f0": 72,
+    "?": 73,
+    "\u2019": 74,
+    "\u2018": 75,
+    "\ud83c\udfe0": 76,
+    "\u2014": 77,
+    "\ud83c\udfe1": 78,
+    "\u2728": 79,
+    "\u00ae": 80,
+    "\ud83c\udf38": 81,
+    "$": 82,
+    "\ud83d\udc8c": 83,
+    "\ud83d\udcb8": 84,
+    "\ud83d\udd52": 85,
+    "\ud83d\udfe2": 86,
+    "\ud83d\ude97": 87,
+    "\ud83e\udde0": 88,
+    "\ud83d\udc64": 89,
+    "\ud83c\udf89": 90,
+    "\ud83d\ude80": 91,
+    "\ud83c\udf0e": 92,
+    "\ud83d\udc40": 93,
+    "\ufe0f": 94,
+    "\u2011": 95,
+    "\ud83c\udf82": 96,
+    "\ud83d\ude08": 97,
+    "\ud83c\udfa4": 98,
+    ";": 99,
+    "\ud83d\udcbc": 100,
+    "\ud83e\udd76": 101,
+    "\ud83d\udea8": 102,
+    "\ud83c\udf34": 103,
+    "\ud83d\udd2d": 104,
+    "\ud83d\ude0e": 105,
+    "\u2600": 106,
+    "\ud83d\udcda": 107,
+    "\ud83c\udf81": 108,
+    "\ud83e\udd16": 109,
+    "\ud83d\udc4b": 110,
+    "\u2022": 111,
+    "\u2763": 112,
+    "\ud83d\udd25": 113,
+    "\ud83c\udf40": 114,
+    "\u2764": 115,
+    "\u200d": 116,
+    "\"": 117,
+    "\ud83d\udcf1": 118
+  },
+  "next_id": 119
+}