ai-tech-articles / index.py
siavava's picture
update load script
b20e4b8 verified
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Simple script to generate metadata about corpus.
"""
__author__ = "Amittai Siavava"
__version__ = "0.0.1"
from os import mkdir
from collections import Counter
import csv
import pandas as pd
def index_pages():
"""
Generate a friendly index of the pages.
We create a csv and a tsv (in case one proves more convenient than the other).
"""
docID = 0
with open("raw2.csv", "w") as csv_file:
writer = csv.writer(csv_file)
writer.writerow(["id", "year", "title", "url", "text"])
while True:
try:
with open(f"../data/log/{docID}", "r") as meta, open(f"../data/log/{docID}.txt", "r") as data:
title = meta.readline().strip()
year = meta.readline().strip()
# if year starts with '|', append to title and read next line
if year.startswith("|"):
title += " " + year
year = meta.readline().strip()
url = meta.readline().strip()
text = data.read()
meta.close()
data.close()
if year and 2000 <= int(year) <= 2023:
print(f"Indexing: {docID}")
writer.writerow([docID, year, f'"{title}"', f'"{url}"', f'"{text}"'])
docID += 1
except Exception as e:
print(e)
print(f"{docID = }")
break
print("Done.")
print(f"okay...")
# save to file
# df = pd.read_csv("raw.csv")
print("Done.")
def categorize():
"""
Categorize the pages by year.
"""
docID = 0
years = Counter()
years_index = { str(year) for year in range(2000, 2024)}
while True:
try:
with open(f"../log/{docID}.txt", "r") as doc, open(f"../log/{docID}", "r") as meta:
title = meta.readline().strip()
year = meta.readline().strip()
url = meta.readline().strip()
text = doc.read()
doc.close()
meta.close()
if year in years_index:
try:
mkdir(f"../categorized/{year}")
except:
pass
id = years[year]
with open(f"../categorized/{year}/{id}.txt", "w") as f:
f.write(f"{title}\n{year}\n{url}\n\n{text}")
f.close()
years[year] += 1
docID += 1
except:
break
if __name__ == "__main__":
index_pages()
# categorize()
# load_data()