refactor: replace pandas with csv module for timezone data loading

test_print_timezones.py

@@ -1,6 +1,6 @@
 import pytz
 from datetime import datetime
-import pandas as pd
+from csv import DictReader
 
 
 def get_tz_info(tz_name: str, timezones: list[dict]) -> dict | None:
@@ -41,52 +41,74 @@ def load_tz_file():
                       "abbreviation", "time_start", "gmt_offset", "dst"]
     # columns to load
     load_columns = ["zone_name", "country_code"]
-    # read csv with pandas
-    df = pd.read_csv(timezone_file, names=timezone_names)
+    # read csv
+    with open(timezone_file, newline='') as csvfile:
+        reader = DictReader(csvfile, fieldnames=timezone_names)
+        csv = [row for row in reader]
     # drop all columns except load_columns
-    df = df[load_columns]
+    csv = [{k: v for k, v in row.items() if k in load_columns} for row in csv]
     # distinct zone_names
-    df = df.drop_duplicates(subset=["zone_name"])
+    seen = set()
+    unique_csv = []
+    for row in csv:
+        if row["zone_name"] not in seen:
+            seen.add(row["zone_name"])
+            unique_csv.append(row)
+    csv = unique_csv
 
-    # reset index
-    df = df.reset_index(drop=True)
-
-    return df
+    return csv
 
 
 def main():
 
-    # read csv with pandas
-    df_file = load_tz_file()
+    # read csv file and load timezones and countries
+    csv = load_tz_file()
 
     # split zone_name into components by "/"
-    df_file[['region', 'city']] = df_file['zone_name'].str.split(
-        '/', expand=True, n=1)
+    for row in csv:
+        parts = row["zone_name"].split("/", 1)
+        row["region"] = parts[0]
+        row["city"] = parts[1] if len(parts) > 1 else None
     # drop regions with no country_code (like Etc, GMT, etc)
-    df_file = df_file[df_file['country_code'].notna()]
+    csv = [row for row in csv if row["country_code"]]
+
+    # get all timezones from pytz and split into region and city
+
+    tz = [{"zone_name": tz} for tz in pytz.all_timezones]
 
-    df_tz = pd.DataFrame(pytz.all_timezones)
-    # rename column to zone_name
-    df_tz = df_tz.rename(columns={0: 'zone_name'})
     # split zone_name into components by "/"
-    df_tz[['region', 'city']] = df_tz['zone_name'].str.split(
-        '/', expand=True, n=1)
+    for row in tz:
+        parts = row["zone_name"].split("/", 1)
+        row["region"] = parts[0]
+        row["city"] = parts[1] if len(parts) > 1 else None
     # drop regions with no city (like UTC, GMT, etc)
-    df_tz = df_tz[df_tz['city'].notna()]
+    tz = [row for row in tz if row["city"]]
     # drop rows where region is 'Etc'
-    df_tz = df_tz[df_tz['region'] != 'Etc']
+    tz = [row for row in tz if row["region"] != "Etc"]
+
+    # join data on region and city
+    timezones = []
+    for tz_row in tz:
+        for csv_row in csv:
+            if tz_row["region"] == csv_row["region"] and tz_row["city"] == csv_row["city"]:
+                timezones.append({
+                    "zone_name": tz_row["zone_name"],
+                    "country_code": csv_row["country_code"],
+                    "region": tz_row["region"],
+                    "city": tz_row["city"],
+                })
+                break
 
-    # join dataframes on region and city
-    df_merged = pd.merge(df_file, df_tz, on=[
-                         'region', 'city'], how='inner', indicator=True)
     # reorder columns
-    df_merged = df_merged[['region', 'city', 'country_code']]
-    # print merged dataframe
-    print(f"Merged timezones: {len(df_merged)}")
-    print(df_merged.sample(20).to_string(index=False))
-    regions = df_merged['region'].unique()
+    timezones = [{k: row[k] for k in ['region', 'city', 'country_code']}
+                 for row in timezones]
+
+    # print merged data
+    print(f"Merged timezones: {len(timezones)}")
+    print(timezones[:20])
+    regions = set(row['region'] for row in timezones)
     for region in regions:
-        df_region = df_merged[df_merged['region'] == region]
+        df_region = [row for row in timezones if row['region'] == region]
         print(f"{len(df_region)} merged in {region}")