Add vectorized Epstein database with LFS

.gitattributes

@@ -57,3 +57,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+*.faiss filter=lfs diff=lfs merge=lfs -text
+*.jsonl filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,3 +1,35 @@
----
-license: mit
----
+# Epstein Vectorized Database
+
+This repository contains a vectorized document database built from a corpus of text documents. The database consists of:
+
+- `epstein_index.faiss`: FAISS index of dense vector embeddings (384-dim, all-MiniLM-L6-v2, cosine similarity)
+- `epstein_metadata.parquet`: Parquet file with metadata for each vector (id, filename, text_snippet, chars, words)
+- `epstein_metadata.jsonl`: Same metadata as JSONL (one object per line)
+
+## Usage
+
+- Load the FAISS index and metadata in Python:
+  ```python
+  import faiss
+  import pandas as pd
+  index = faiss.read_index('epstein_index.faiss')
+  meta = pd.read_parquet('epstein_metadata.parquet')
+  ```
+- Encode a query with the same model (`all-MiniLM-L6-v2` from sentence-transformers), normalize, and search:
+  ```python
+  from sentence_transformers import SentenceTransformer
+  import numpy as np
+  model = SentenceTransformer('all-MiniLM-L6-v2')
+  q_emb = model.encode(['your query'], convert_to_numpy=True)
+  q_emb = q_emb / np.linalg.norm(q_emb, axis=1, keepdims=True)
+  D, I = index.search(q_emb.astype('float32'), k=5)
+  results = meta.iloc[I[0]]
+  ```
+
+## Files
+- `epstein_index.faiss`: FAISS index (IndexFlatIP, 384-dim, normalized vectors)
+- `epstein_metadata.parquet`: Parquet metadata (id, filename, text_snippet, chars, words)
+- `epstein_metadata.jsonl`: JSONL metadata (same fields)
+
+## License
+No original EPS file content is included. This database contains only vectorized representations and metadata.

dataset_card.json

@@ -0,0 +1,16 @@
+{
+  "dataset_name": "epstein-vectorized-db",
+  "description": "A vectorized document database with FAISS index and metadata, built from a corpus of text documents. No original EPS file content is included.",
+  "features": {
+    "faiss_index": "FAISS IndexFlatIP, 384-dim, all-MiniLM-L6-v2, cosine similarity",
+    "metadata_parquet": "Parquet file with id, filename, text_snippet, chars, words",
+    "metadata_jsonl": "JSONL file with same fields as Parquet"
+  },
+  "usage": "Load the FAISS index and metadata in Python. Encode queries with all-MiniLM-L6-v2, normalize, and search for similar documents.",
+  "license": "No original EPS file content is included. This database contains only vectorized representations and metadata.",
+  "files": [
+    "epstein_index.faiss",
+    "epstein_metadata.parquet",
+    "epstein_metadata.jsonl"
+  ]
+}

epstein_index.faiss

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e887e34faab60dea926d974784044ca4688669e733e2386bc48de4b4558a277
+size 39628845

epstein_metadata.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c463494e6ca7b6f90c3e14f772da0109c7a9e4ea88991480c5ae99ce0898ad2d
+size 15292705

epstein_metadata.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4457345948898486526266e973820bfe95ead291f526f9e1be2333717dd4d657
+size 8567248