CalvinBackup/OpenAirframes
1
0
Fork 0
mirror of https://github.com/PlaneQuery/OpenAirframes.git synced 2026-07-20 16:50:49 +02:00
Code Issues Packages Projects Releases Wiki Activity

stop depue that destroys previous days

Browse Source
This commit is contained in:
ggman12
2026-02-15 17:55:16 -05:00
parent 1a553d5f44
commit 3f38263a0c
1 changed files with 37 additions and 55 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/adsb/combine_chunks_to_csv.py
+1 -19
View File
@@ -93,13 +93,11 @@ def download_and_merge_base_release(compressed_df: pl.DataFrame) -> tuple[pl.Dat
from src.get_latest_release import download_latest_aircraft_adsb_csv from src.get_latest_release import download_latest_aircraft_adsb_csv
print("Downloading base ADS-B release...") print("Downloading base ADS-B release...")
try:
base_path = download_latest_aircraft_adsb_csv( base_path = download_latest_aircraft_adsb_csv(
output_dir="./data/openairframes_base" output_dir="./data/openairframes_base"
) )
print(f"Download returned: {base_path}") print(f"Download returned: {base_path}")
if base_path and os.path.exists(str(base_path)):
print(f"Loading base release from {base_path}") print(f"Loading base release from {base_path}")
# Extract start date from filename (e.g., openairframes_adsb_2025-05-01_2026-02-14.csv.gz) # Extract start date from filename (e.g., openairframes_adsb_2025-05-01_2026-02-14.csv.gz)
@@ -132,23 +130,7 @@ def download_and_merge_base_release(compressed_df: pl.DataFrame) -> tuple[pl.Dat
combined = pl.concat([base_df, compressed_df]) combined = pl.concat([base_df, compressed_df])
print(f"After concat: {len(combined)} records") print(f"After concat: {len(combined)} records")
deduplicated = combined.unique(subset=["icao"], keep="last") return combined, earliest_date
print(f"Combined with base: {len(combined)} -> {len(deduplicated)} after dedup")
del base_df, combined
gc.collect()
return deduplicated, earliest_date
else:
print(f"No base release found at {base_path}, using only new data")
return compressed_df, None
except Exception as e:
import traceback
print(f"Failed to download base release: {e}")
traceback.print_exc()
return compressed_df, None
def cleanup_chunks(output_id: str, chunks_dir: str): def cleanup_chunks(output_id: str, chunks_dir: str):
"""Delete chunk parquet files after successful merge.""" """Delete chunk parquet files after successful merge."""