File size: 10,935 Bytes
3609368 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 | import os
import re
from tqdm import tqdm
import pandas as pd
import config
def load_crosswalks():
"""Loads NAICS->SIC, SIC->Ticker, and Legislator Mappings."""
print("Loading Crosswalks...")
naics_sic = pd.read_csv(config.NAICS_TO_SIC_PATH)
naics_sic['NAICS'] = naics_sic['NAICS'].astype(str).str.replace(r'\.0$', '', regex=True).str.zfill(6)
naics_sic['SIC'] = naics_sic['SIC'].astype(str).str.replace(r'\.0$', '', regex=True).str.zfill(4)
naics_to_sic_map = naics_sic.groupby('NAICS')['SIC'].apply(list).to_dict()
if os.path.exists(config.COMPANY_SIC_DATA_PATH):
sic_data = pd.read_csv(config.COMPANY_SIC_DATA_PATH)
sic_data['sic'] = sic_data['sic'].astype(str).str.replace(r'\.0$', '', regex=True).str.zfill(4)
sic_to_ticker_map = sic_data.groupby('sic')['ticker'].apply(list).to_dict()
else:
print(f"WARNING: {config.COMPANY_SIC_DATA_PATH} not found.")
sic_to_ticker_map = {}
if os.path.exists(config.LEGISLATORS_CROSSWALK_PATH):
leg_df = pd.read_csv(config.LEGISLATORS_CROSSWALK_PATH, low_memory=False)
leg_map_df = leg_df[['id_icpsr', 'id_bioguide']].dropna().drop_duplicates()
leg_map_df['id_icpsr'] = leg_map_df['id_icpsr'].astype(int)
icpsr_to_bioguide = leg_map_df.set_index('id_icpsr')['id_bioguide'].to_dict()
else:
print(f"WARNING: Legislator crosswalk not found at {config.LEGISLATORS_CROSSWALK_PATH}. Voting edges will fail.")
icpsr_to_bioguide = {}
return naics_to_sic_map, sic_to_ticker_map, icpsr_to_bioguide
def parse_bill_ids(id_str):
try:
if pd.isna(id_str): return []
matches = re.findall(r'[a-zA-Z0-9]+-[0-9]+', str(id_str))
return [m.lower().strip() for m in matches]
except:
return []
def build_lobbying_events():
print("Building Lobbying & Voting Event Stream (Using Filing Dates)...")
bills_df = pd.read_csv(config.LOBBYING_BILLS_PATH)
clients_df = pd.read_csv(config.LOBBYING_CLIENTS_PATH)
reports_df = pd.read_csv(config.LOBBYING_REPORTS_PATH)
issues_df = pd.read_csv(config.LOBBYING_ISSUES_PATH)
has_votes = os.path.exists(config.VOTEVIEW_VOTES_PATH) and os.path.exists(config.VOTEVIEW_ROLLCALLS_PATH)
if has_votes:
print("Loading VoteView Data...")
votes_df = pd.read_csv(config.VOTEVIEW_VOTES_PATH)
rollcalls_df = pd.read_csv(config.VOTEVIEW_ROLLCALLS_PATH, low_memory=False)
bills_df['bill_id'] = bills_df['bill_id'].astype(str).str.lower().str.strip()
reports_df['report_uuid'] = reports_df['report_uuid'].astype(str)
reports_df['lob_id'] = reports_df['lob_id'].astype(str)
clients_df['lob_id'] = clients_df['lob_id'].astype(str)
naics_map, sic_ticker_map, icpsr_map = load_crosswalks()
# --- Step 1: Map Clients to Tickers ---
clients_df['naics_str'] = clients_df['naics'].astype(str).str.replace(r'\.0$', '', regex=True).str.zfill(6)
client_ticker_records = []
print("Mapping Clients to Tickers...")
for idx, row in tqdm(clients_df.iterrows(), total=len(clients_df), desc="Mapping Clients"):
target_tickers = []
for sic in naics_map.get(row['naics_str'], []):
target_tickers.extend(sic_ticker_map.get(sic, []))
if target_tickers:
for t in set(target_tickers):
client_ticker_records.append({'lob_id': row['lob_id'], 'ticker': t})
client_ticker_df = pd.DataFrame(client_ticker_records)
print(f"Mapped {client_ticker_df['lob_id'].nunique()} clients to {client_ticker_df['ticker'].nunique()} unique tickers.")
if client_ticker_df.empty:
print("No clients mapped to tickers. Exiting.")
return
# --- Step 2: Link Reports to Bills ---
print("Parsing Bill IDs from Issues...")
tqdm.pandas(desc="Parsing Bill IDs")
issues_with_bills = issues_df.dropna(subset=['bill_id_agg']).copy()
issues_with_bills['bill_id_list'] = issues_with_bills['bill_id_agg'].progress_apply(parse_bill_ids)
issues_exploded = issues_with_bills.explode('bill_id_list').rename(columns={'bill_id_list': 'bill_id'})
issues_exploded['report_uuid'] = issues_exploded['report_uuid'].astype(str)
issues_exploded['bill_id'] = issues_exploded['bill_id'].astype(str)
print("Merging Issues with Reports...")
bill_client_chain = pd.merge(
issues_exploded[['bill_id', 'report_uuid']],
reports_df[['report_uuid', 'lob_id', 'estimated_filing_date']],
on='report_uuid',
how='inner'
)
# NEW OPTIMIZATION: Drop duplicates BEFORE merging to save massive amounts of memory
base_chain = bill_client_chain[['lob_id', 'bill_id', 'estimated_filing_date']].drop_duplicates()
ticker_bill_df = pd.merge(base_chain, client_ticker_df, on='lob_id').drop(columns=['lob_id'])
ticker_bill_df = ticker_bill_df.drop_duplicates()
all_events_dfs = []
# --- Step 3: Strong Edges (Sponsorship) ---
if getattr(config, 'INCLUDE_LOBBYING_SPONSORSHIP', True):
print("Generating Strong Edges (Sponsorship)...")
strong_df = pd.merge(ticker_bill_df, bills_df[['bill_id', 'bioguide_id']], on='bill_id', how='inner')
strong_df = strong_df.dropna(subset=['bioguide_id', 'estimated_filing_date'])
# Vectorized Event Creation (Replaces the 9-minute loop)
strong_df = strong_df[['estimated_filing_date', 'ticker', 'bioguide_id']].drop_duplicates()
strong_df.rename(columns={'estimated_filing_date': 'date'}, inplace=True)
strong_df['event_type'] = 'LOBBY_STRONG'
strong_df['weight'] = 1.0
print(f"Valid Sponsorship Connections Found: {len(strong_df)}")
all_events_dfs.append(strong_df)
else:
print("Skipping Strong Edges (Sponsorship) per config.")
# --- Step 4: Weak Edges (Voting) ---
if getattr(config, 'INCLUDE_LOBBYING_VOTING', True) and has_votes:
print("Generating Weak Edges (Voting)...")
if 'bill_number' in bills_df.columns and 'bill_number' in rollcalls_df.columns:
lobbied_bills = ticker_bill_df['bill_id'].unique()
target_bills_df = bills_df[bills_df['bill_id'].isin(lobbied_bills)][['bill_id', 'bill_type', 'bill_number', 'congress_number']].copy()
target_bills_df['clean_type'] = target_bills_df['bill_type'].astype(str).str.replace(r'[^a-zA-Z]', '', regex=True).str.upper()
target_bills_df['clean_num'] = target_bills_df['bill_number'].astype(str).str.replace(r'\.0$', '', regex=True).str.strip()
target_bills_df['vv_bill_number'] = target_bills_df['clean_type'] + target_bills_df['clean_num']
target_bills_df['congress_number'] = target_bills_df['congress_number'].astype(str).str.replace(r'\.0$', '', regex=True).str.strip()
rollcalls_merge = rollcalls_df[['congress', 'rollnumber', 'bill_number']].copy()
rollcalls_merge['vv_bill_number'] = rollcalls_merge['bill_number'].astype(str).str.replace(r'[^a-zA-Z0-9]', '', regex=True).str.upper()
rollcalls_merge['congress'] = rollcalls_merge['congress'].astype(str).str.replace(r'\.0$', '', regex=True).str.strip()
rollcalls_merge['rollnumber'] = rollcalls_merge['rollnumber'].astype(str).str.replace(r'\.0$', '', regex=True).str.strip()
bill_votes_map = pd.merge(
target_bills_df[['bill_id', 'congress_number', 'vv_bill_number']],
rollcalls_merge[['congress', 'rollnumber', 'vv_bill_number']],
left_on=['congress_number', 'vv_bill_number'],
right_on=['congress', 'vv_bill_number']
)
bill_votes_map = bill_votes_map[~bill_votes_map['congress'].isin(['nan', 'None', ''])]
bill_votes_map = bill_votes_map[~bill_votes_map['rollnumber'].isin(['nan', 'None', ''])]
yea_votes = votes_df[votes_df['cast_code'].isin([1, 2, 3])].copy()
yea_votes['bioguide_id'] = yea_votes['icpsr'].map(icpsr_map)
yea_votes = yea_votes.dropna(subset=['bioguide_id', 'congress', 'rollnumber'])
yea_votes['congress'] = yea_votes['congress'].astype(str).str.replace(r'\.0$', '', regex=True).str.strip()
yea_votes['rollnumber'] = yea_votes['rollnumber'].astype(str).str.replace(r'\.0$', '', regex=True).str.strip()
yea_votes = yea_votes[~yea_votes['congress'].isin(['nan', 'None', ''])]
print("Mapping Legislator votes to Bills...")
bill_to_legislator = pd.merge(
bill_votes_map[['bill_id', 'congress', 'rollnumber']],
yea_votes[['congress', 'rollnumber', 'bioguide_id']],
on=['congress', 'rollnumber']
)
bill_to_legislator = bill_to_legislator[['bill_id', 'bioguide_id']].drop_duplicates()
print("Merging Voting Records with Lobbying Clients (Chunked to prevent Memory Error)...")
# NEW OPTIMIZATION: Process one Ticker at a time to prevent Cartesian explosion
weak_events_list = []
valid_bills_with_votes = bill_to_legislator['bill_id'].unique()
tb_subset = ticker_bill_df[ticker_bill_df['bill_id'].isin(valid_bills_with_votes)]
for ticker, group in tqdm(tb_subset.groupby('ticker'), desc="Building Weak Edges"):
merged = pd.merge(group[['bill_id', 'estimated_filing_date']], bill_to_legislator, on='bill_id')
unique_edges = merged[['estimated_filing_date', 'bioguide_id']].drop_duplicates()
unique_edges['ticker'] = ticker
weak_events_list.append(unique_edges)
if weak_events_list:
weak_df = pd.concat(weak_events_list, ignore_index=True)
weak_df.rename(columns={'estimated_filing_date': 'date'}, inplace=True)
weak_df['event_type'] = 'LOBBY_WEAK'
weak_df['weight'] = 0.5
print(f"Valid Voting Connections Found: {len(weak_df)}")
all_events_dfs.append(weak_df)
else:
print("Valid Voting Connections Found: 0")
else:
print("Skipping Weak Edges (Voting) per config.")
if all_events_dfs:
final_events_df = pd.concat(all_events_dfs, ignore_index=True)
print(f"\nSUCCESS: Generated {len(final_events_df)} total Lobbying/Voting Events.")
final_events_df.to_csv(config.LOBBYING_EVENTS_PATH, index=False)
else:
print("\nGenerated 0 total events.")
if __name__ == "__main__":
build_lobbying_events() |