handle duplictes much better

.github/workflows/planequery-aircraft-daily-release.yaml

@@ -39,7 +39,7 @@ jobs:
         id: meta
         run: |
           DATE=$(date -u +"%Y-%m-%d")
-          TAG="faa-${DATE}"
+          TAG="planequery-aircraft-${DATE}"
           # Find the CSV file in data/planequery_aircraft matching the pattern
           CSV_FILE=$(ls data/planequery_aircraft/planequery_aircraft_*_${DATE}.csv | head -1)
           CSV_BASENAME=$(basename "$CSV_FILE")

src/create_daily_planequery_aircraft_release.py

@@ -7,20 +7,20 @@ out_dir.mkdir(parents=True, exist_ok=True)
 zip_name = f"ReleasableAircraft_{date_str}.zip"
 
 zip_path = out_dir / zip_name
+if not zip_path.exists():
+    # URL and paths
+    url = "https://registry.faa.gov/database/ReleasableAircraft.zip"
+    from urllib.request import Request, urlopen
 
-# URL and paths
+    req = Request(
-url = "https://registry.faa.gov/database/ReleasableAircraft.zip"
+        url,
-from urllib.request import Request, urlopen
+        headers={"User-Agent": "Mozilla/5.0"},
+        method="GET",
+    )
 
-req = Request(
+    with urlopen(req, timeout=120) as r:
-    url,
+        body = r.read()
-    headers={"User-Agent": "Mozilla/5.0"},
+        zip_path.write_bytes(body)
-    method="GET",
-)
-
-with urlopen(req, timeout=120) as r:
-    body = r.read()
-    zip_path.write_bytes(body)
 
 OUT_ROOT = Path("data/planequery_aircraft")
 OUT_ROOT.mkdir(parents=True, exist_ok=True)

src/derive_from_faa_master_txt.py

@@ -66,7 +66,7 @@ def normalize(s: pd.Series) -> pd.Series:
 
 
 def concat_faa_historical_df(df_base, df_new):
-
+    df_new = df_new[df_base.columns]
     df_base = pd.concat([df_base, df_new], ignore_index=True)
     
     CONTENT_COLS = [
@@ -74,10 +74,49 @@ def concat_faa_historical_df(df_base, df_new):
         if c not in {"download_date"}
     ]
     
+    # Normalize values to handle numeric type, formatting, and list ordering differences
+    def normalize_series(series):
+        def normalize_value(val):
+            # Handle lists (sort them for consistent comparison)
+            if isinstance(val, list):
+                return "|".join(sorted(str(v) for v in val))
+            
+            # Convert to string
+            val_str = str(val).strip()
+            
+            # Handle empty strings
+            if val_str == "" or val_str == "nan":
+                return ""
+            
+            # Check if it looks like a list representation (starts with [ )
+            if val_str.startswith('[') and val_str.endswith(']'):
+                try:
+                    # Try to parse as a list-like string
+                    import ast
+                    parsed = ast.literal_eval(val_str)
+                    if isinstance(parsed, list):
+                        return "|".join(sorted(str(v) for v in parsed))
+                except (ValueError, SyntaxError):
+                    pass  # Not a valid list, continue to other checks
+            
+            # Try to normalize as number
+            try:
+                # Remove leading zeros and convert float/int representations
+                num_val = float(val_str)
+                # If it's a whole number, return as int string (no .0)
+                if num_val == int(num_val):
+                    return str(int(num_val))
+                # Otherwise return as float
+                return str(num_val)
+            except (ValueError, OverflowError):
+                # Not a number, return as-is
+                return val_str
+        
+        return series.apply(normalize_value)
+    
     df_base["row_fingerprint"] = (
         df_base[CONTENT_COLS]
-        .fillna("")
+        .apply(normalize_series, axis=0)
-        .astype(str)
         .apply(lambda row: "|".join(row), axis=1)
     )
     

src/get_latest_planequery_aircraft_release.py

@@ -126,12 +126,13 @@ def download_latest_aircraft_csv(
     return saved_to
 
 def get_latest_aircraft_csv_df():
-    csv_path = download_latest_aircraft_csv()
+    # csv_path = download_latest_aircraft_csv()
+    csv_path = '/Users/jonahgoode/Documents/PlaneQuery/Code/planequery-aircraft/data/planequery_aircraft/planequery_aircraft_2023-08-16_2026-01-31.csv'
     import pandas as pd
     df = pd.read_csv(csv_path, dtype={'transponder_code': str, 
            'unique_regulatory_id': str, 
            'registrant_county': str})
-    
+    df = df.fillna("")
     # Extract date from filename pattern: planequery_aircraft_{date}_{date}.csv
     match = re.search(r"planequery_aircraft_(\d{4}-\d{2}-\d{2})_", str(csv_path))
     if not match: