Upload scripts/form4_add_roles.py with huggingface_hub

scripts/form4_add_roles.py

@@ -0,0 +1,120 @@
+# form4_add_roles.py
+# Adds is_officer, is_director, is_ten_pct_owner, officer_title columns
+# to the existing form4_transactions.csv.
+#
+# The SEC's new REPORTINGOWNER format encodes roles as a comma-separated
+# string in rptowner_relationship, e.g. "Director,Officer,TenPercentOwner"
+# and officer title in rptowner_title.
+#
+# Re-downloads all 81 quarterly ZIPs (REPORTINGOWNER table only — small),
+# merges onto form4_transactions.csv, derives boolean flags, saves.
+#
+# Run: python scripts/form4_add_roles.py
+
+import io
+import time
+import zipfile
+import requests
+import pandas as pd
+from tqdm import tqdm
+
+OUT_PATH = r"D:\UoE AI\Dissertation\IPP Draft\datasets\form4_transactions.csv"
+HEADERS  = {'User-Agent': 'S2880814 University of Edinburgh s.g.vishnu@sms.ed.ac.uk'}
+SLEEP    = 0.5
+TIMEOUT  = 120
+
+# ── 1. Load existing clean file ────────────────────────────────────────────────
+print("Loading form4_transactions.csv...")
+df = pd.read_csv(OUT_PATH, low_memory=False)
+print(f"  {len(df):,} rows, {len(df.columns)} columns")
+our_accessions = set(df['accession_number'].dropna())
+
+# ── 2. Re-download REPORTINGOWNER from all 81 quarters ────────────────────────
+def quarter_urls():
+    urls, year, qtr = [], 2006, 1
+    while (year, qtr) <= (2026, 1):
+        urls.append((year, qtr,
+            f"https://www.sec.gov/files/structureddata/data/insider-transactions-data-sets/{year}q{qtr}_form345.zip"))
+        qtr += 1
+        if qtr > 4:
+            qtr, year = 1, year + 1
+    return urls
+
+print(f"\nFetching REPORTINGOWNER from 81 quarters for {len(our_accessions):,} accessions...")
+owner_frames = []
+
+for year, qtr, url in tqdm(quarter_urls(), desc="Quarters"):
+    time.sleep(SLEEP)
+    try:
+        r = requests.get(url, headers=HEADERS, timeout=TIMEOUT)
+        if r.status_code != 200:
+            continue
+        zf       = zipfile.ZipFile(io.BytesIO(r.content))
+        own_file = next((n for n in zf.namelist() if 'reportingowner' in n.lower()), None)
+        if not own_file:
+            continue
+        own = pd.read_csv(io.BytesIO(zf.read(own_file)), sep='\t',
+                          low_memory=False, on_bad_lines='skip')
+        own.columns = own.columns.str.lower().str.strip()
+
+        own = own[own['accession_number'].isin(our_accessions)]
+        if own.empty:
+            continue
+
+        keep = [c for c in ['accession_number', 'rptownercik',
+                             'rptowner_relationship', 'rptowner_title'] if c in own.columns]
+        owner_frames.append(own[keep].copy())
+
+    except Exception as e:
+        tqdm.write(f"  [WARN] {year} Q{qtr}: {e}")
+
+# ── 3. Build roles lookup table ────────────────────────────────────────────────
+print("\nBuilding roles lookup...")
+owner_df = pd.concat(owner_frames, ignore_index=True)
+owner_df = owner_df.drop_duplicates(subset=['accession_number', 'rptownercik'])
+
+# Parse rptowner_relationship → boolean flags
+rel = owner_df['rptowner_relationship'].fillna('')
+owner_df['is_officer']       = rel.str.contains('Officer',          case=False).astype(int)
+owner_df['is_director']      = rel.str.contains('Director',         case=False).astype(int)
+owner_df['is_ten_pct_owner'] = rel.str.contains('TenPercentOwner',  case=False).astype(int)
+
+if 'rptowner_title' in owner_df.columns:
+    owner_df.rename(columns={'rptowner_title': 'officer_title'}, inplace=True)
+
+owner_df.drop(columns=['rptowner_relationship'], inplace=True)
+print(f"  Owner rows: {len(owner_df):,}")
+print(f"  Officers  : {owner_df['is_officer'].sum():,}")
+print(f"  Directors : {owner_df['is_director'].sum():,}")
+
+# ── 4. Merge onto main dataframe ───────────────────────────────────────────────
+print("\nMerging roles onto transactions...")
+
+# Drop any stale role columns if re-running
+for col in ['is_officer', 'is_director', 'is_ten_pct_owner', 'officer_title']:
+    if col in df.columns:
+        df.drop(columns=[col], inplace=True)
+
+df = df.merge(owner_df, on=['accession_number', 'rptownercik'], how='left')
+
+# Move role cols next to owner_name
+cols = df.columns.tolist()
+role_cols = [c for c in ['is_officer', 'is_director', 'is_ten_pct_owner', 'officer_title'] if c in cols]
+for rc in reversed(role_cols):
+    cols.remove(rc)
+    insert_at = cols.index('rptownercik') + 1
+    cols.insert(insert_at, rc)
+df = df[cols]
+
+# ── 5. Save ────────────────────────────────────────────────────────────────────
+df.to_csv(OUT_PATH, index=False)
+
+print(f"\n{'='*60}")
+print(f"  Final shape  : {df.shape}")
+print(f"  Columns      : {df.columns.tolist()}")
+print(f"\nSample:")
+preview = [c for c in ['ticker', 'owner_name', 'officer_title', 'is_officer',
+                        'is_director', 'transaction_date', 'transaction_code',
+                        'acquired_disposed', 'shares', 'price_per_share'] if c in df.columns]
+print(df[preview].head(8).to_string(index=False))
+print(f"\nSaved -> {OUT_PATH}")