fio-l1-vienna-shapefile-recovery

Rerun

import geopandas as gpd
import pandas as pd
import json

def migrate_parcels(shapefile_path, column_map_path, output_path):
    # 1. Load column map
    # The CSV looks like it might have some leading whitespace or strange formatting based on the Read output
    # It has: index (skipped?), truncated, original
    # Let's try reading it with pandas and cleaning it
    try:
        # The Read output showed: "     1\ttruncated,original"
        # It looks like a tab-separated or space-separated header followed by comma-separated values.
        # Actually, looking closely at:
        # "     1	truncated,original"
        # "     2	KATASTRALG,KATASTRALGEMEINDE_NAME"
        # It seems it's tab-separated for the first part and comma-separated for the rest? 
        # Or maybe it's just a standard CSV where the first column is index and there's a tab?
        # Let's try reading it as a CSV with whitespace issues.
        df_map = pd.read_csv(column_map_path, sep=None, engine='python')
        print("Column map loaded successfully:")
        print(df_map)
    except Exception as e:
        print(f"Error reading column map: {e}")
        return

    # Clean column map to create a dictionary {truncated: original}
    # Based on Read output:
    # Col 0: Index?
    # Col 1: truncated
    # Col 2: original
    # We need to handle the column names correctly.
    
    # Let's inspect columns of df_map
    print(f"Columns in df_map: {df_map.columns.tolist()}")

    # Depending on how read_csv parsed it, we might need to rename or select
    # Let's assume it has columns like [' ', 'truncated', 'original'] or similar.
    # Let's standardize the mapping.
    mapping_dict = {}
    
    # Find the columns that correspond to truncated and original
    # Based on the content:
    # 2 KATASTRALG,KATASTRALGEMEINDE_NAME
    # It looks like 'truncated' and 'original' are the headers.
    
    # We'll use the values from 'truncated' and 'original' columns
    # We need to handle potential NaNs or incorrect parsing
    for _, row in df_map.iterrows():
        # Using index-based access in case names are weird
        # We want the mapping: row[truncated] -> row[original]
        # In the Read output, 'truncated' and 'original' are headers.
        # Let's try to find them.
        
        # It's safer to find the column names by checking if 'truncated' is in the columns
        trunc_col = None
        orig_col = None
        for col in df_map.columns:
            if 'truncated' in col.lower():
                trunc_col = col
            if 'original' in col.lower():
                orig_col = col
        
        if trunc_col and orig_col:
            mapping_dict[str(row[trunc_col]).strip()] = str(row[orig_col]).strip()
            break # we found the loop logic inside? no, the loop should be outside.
    
    # Redoing the mapping extraction logic properly
    mapping_dict = {}
    trunc_col = None
    orig_col = None
    for col in df_map.columns:
        if 'truncated' in col.lower():
            trunc_col = col
        if 'original' in col.lower():
            orig_col = col
            
    if trunc_col and orig_col:
        for _, row in df_map.iterrows():
            val_trunc = str(row[trunc_col]).strip()
            val_orig = str(row[orig_col]).strip()
            if val_trunc != 'nan' and val_orig != 'nan':
                mapping_dict[val_trunc] = val_orig
    
    print(f"Generated mapping dictionary: {mapping_dict}")

    # 2. Load Shapefile
    print(f"Loading shapefile: {shapefile_path}")
    gdf = gpd.read_file(shapefile_path)
    print(f"Initial columns: {gdf.columns.tolist()}")

    # 3. Rename columns
    # We only rename columns that are in our mapping dictionary
    current_cols = gdf.columns.tolist()
    new_cols = []
    for col in current_cols:
        if col in mapping_dict:
            new_cols.append(mapping_dict[col])
        else:
            new_cols.append(col)
    
    gdf.columns = new_cols
    print(f"Renamed columns: {gdf.columns.tolist()}")

    # 4. Ensure Geometry is Polygon or MultiPolygon
    # GeoPandas usually handles this, but let's be sure.
    # Filter out non-polygon geometries if any (though usually shapefiles are consistent)
    gdf = gdf[gdf.geometry.type.isin(['Polygon', 'MultiPolygon'])]
    print(f"Geometry types present: {gdf.geometry.type.unique()}")

    # 5. Save to GeoJSON
    print(f"Saving to {output_path}")
    # Using encoding='utf-8' for proper character encoding
    gdf.to_file(output_path, driver='GeoJSON', encoding='utf-8')
    print("Successfully saved GeoJSON.")

if __name__ == "__main__":
    migrate_parcels('parcels.shp', 'column_map.csv', 'parcels.geojson')

import geopandas as gpd
import pandas as pd
import json

def migrate_parcels(shapefile_path, column_map_path, output_path):
    # 1. Load column map
    # The CSV might have some strange leading whitespace or formatting.
    # Based on the previous attempt, it worked, but let's be more robust.
    try:
        # Let's read it with pandas and handle the whitespace.
        # Using sep=None with engine='python' handles various delimiters.
        df_map = pd.read_csv(column_map_path, sep=None, engine='python')
    except Exception as e:
        print(f"Error reading column map: {e}")
        return

    # Clean column names in the mapping file itself
    df_map.columns = [c.strip() for c in df_map.columns]
    
    mapping_dict = {}
    # Find columns that resemble 'truncated' and 'original'
    trunc_col = None
    orig_col = None
    for col in df_map.columns:
        if 'truncated' in col.lower():
            trunc_col = col
        if 'original' in col.lower():
            orig_col = col
            
    if trunc_col and orig_col:
        for _, row in df_map.iterrows():
            val_trunc = str(row[trunc_col]).strip()
            val_orig = str(row[orig_col]).strip()
            # Check for non-null/non-nan values
            if val_trunc != 'nan' and val_orig != 'nan' and val_trunc != '':
                mapping_dict[val_trunc] = val_orig
    
    print(f"Generated mapping dictionary: {mapping_dict}")

    # 2. Load Shapefile
    print(f"Loading shapefile: {shapefile_path}")
    gdf = gpd.read_file(shapefile_path)
    
    # 3. Rename columns
    # The instruction asks for 'correct full column names'.
    # We rename them using the map.
    gdf = gdf.rename(columns=mapping_dict)
    print(f"Columns after renaming: {gdf.columns.tolist()}")

    # 4. Ensure Geometry is Polygon or MultiPolygon
    # This is usually true for parcel datasets, but we filter to be safe.
    gdf = gdf[gdf.geometry.type.isin(['Polygon', 'MultiPolygon'])]

    # 5. Save to GeoJSON
    # GeoJSON is standard UTF-8.
    print(f"Saving to {output_path}")
    gdf.to_file(output_path, driver='GeoJSON', encoding='utf-8')
    print("Successfully saved GeoJSON.")

if __name__ == "__main__":
    migrate_parcels('parcels.shp', 'column_map.csv', 'parcels.geojson')