import os
import pandas as pd
import bibtexparser
import folium
from concurrent.futures import ThreadPoolExecutor
from tqdm import tqdm

# Clear the terminal
os.system('cls' if os.name == 'nt' else 'clear')

# Pfade zu den Dateien
geonames_file_path = 'allCountries.txt'
bib_file_path = 'Research/Charité - Universitätsmedizin Berlin/Systematische Literaturrecherche/Literaturverzeichnis.bib'
cleaned_bib_file_path = 'Research/Charité - Universitätsmedizin Berlin/Systematische Literaturrecherche/cleaned_Literaturverzeichnis.bib'

# Spaltennamen laut Geonames README
columns = [
    'geonameid', 'name', 'asciiname', 'alternatenames', 'latitude', 'longitude',
    'feature class', 'feature code', 'country code', 'cc2', 'admin1 code',
    'admin2 code', 'admin3 code', 'admin4 code', 'population', 'elevation', 'dem',
    'timezone', 'modification date'
]

# GeoNames-Daten laden
print("Lade GeoNames-Daten...")
geo_df = pd.read_csv(geonames_file_path, sep='\t', header=None, names=columns, usecols=['name', 'asciiname', 'latitude', 'longitude'])
geo_df.dropna(subset=['latitude', 'longitude'], inplace=True)
print("GeoNames-Daten geladen.")

# BibTeX-Datei laden
print("Lade BibTeX-Datei...")
with open(bib_file_path, encoding='utf-8') as bibtex_file:
    bib_database = bibtexparser.load(bibtex_file)
print("BibTeX-Datei geladen.")

# Ortsnamen extrahieren und bereinigen
print("Extrahiere und bereinige Ortsnamen...")
locations = set()
for entry in bib_database.entries:
    if 'address' in entry:
        for loc in entry['address'].split(';'):
            locations.update(loc.split(','))

cleaned_locations = {loc.strip() for loc in locations}
print("Ortsnamen extrahiert und bereinigt.")

# Geo-Koordinaten zuordnen
def find_coordinates(location):
    match = geo_df[(geo_df['name'].str.lower() == location.lower()) | (geo_df['asciiname'].str.lower() == location.lower())]
    if not match.empty:
        return match.iloc[0]['latitude'], match.iloc[0]['longitude']
    return None, None

print("Suche Geo-Koordinaten...")
location_coords = {}

for location in tqdm(cleaned_locations, desc="Bearbeitung der Ortsnamen"):
    latitude, longitude = find_coordinates(location)
    if latitude is not None and longitude is not None:
        location_coords[location] = (latitude, longitude)

print("Geo-Koordinaten gefunden.")

# Erstelle die Karte
print("Erstelle Karte...")
map_center = [geo_df['latitude'].mean(), geo_df['longitude'].mean()]
map = folium.Map(location=map_center, zoom_start=2)

for location, coords in location_coords.items():
    folium.Marker(location=coords, popup=location).add_to(map)

map_file_path = 'Research/Charité - Universitätsmedizin Berlin/Systematische Literaturrecherche/literature_map.html'
map.save(map_file_path)
print(f"Karte gespeichert unter {map_file_path}")

# Bereinigen und Speichern der BibTeX-Datei
print("Speichere bereinigte BibTeX-Datei...")
with open(cleaned_bib_file_path, 'w', encoding='utf-8') as bibtex_file:
    bibtexparser.dump(bib_database, bibtex_file)
print(f"Bereinigte BibTeX-Datei gespeichert unter {cleaned_bib_file_path}")