Search for dates and times across all files in a folder.

search-dates.py

"""
Search for dates and times in a folder.

$ python3 search-dates.py facebook-kcimc/
$ python3 search-dates.py Takeout/
"""
import os
import sys
import re
import shutil 
from pathlib import Path
from itertools import islice

root = sys.argv[1] # folder to search

ignore_keywords = ['messages'] # keywords to ignore in directories and filenames
valid_extensions = ['.txt', '.html', '.json', '.csv'] # extensions to search
min_results = 5 # how many matches per file before printing out examples
max_results = 1000 # how many matches per file before we stop counting
printed_results = 10 # how many example matches to print out
terminal_width = shutil.get_terminal_size((80, 20)).columns
padding = terminal_width // 2

# find:
# - timestamps in seconds or milliseconds
# - years 2018-2020
# - time of day
pattern = \
    r'\b15\d{8,11}\b|'+\
    r'\b20(18|19|20)\b|'+\
    r'\d\d?:\d\d:\d\d'

def yellow(e):
    return f'\033[93m{e}\033[00m'

def cyan(e):
    return f'\033[96m{e}\033[00m'

def highlight_matches(e, color):
    parts = []
    prev = 0
    for match in re.finditer(pattern, e):
        start = match.start()
        end = match.end()
        parts.append(e[prev:start])
        parts.append(color(e[start:end]))
        prev = end
    parts.append(e[prev:])
    return ''.join(parts)

for path in Path(root).rglob('*'):
    if not path.is_file():
        continue

    if any([e in str(path) for e in ignore_keywords]):
        continue

    ext = os.path.splitext(path)[1]
    if ext not in valid_extensions:
        continue

    with open(path) as f:
        raw = f.read()
    results = []
    match_iter = re.finditer(pattern, raw)
    for match in islice(match_iter, max_results):
        start = max(0, match.start() - padding)
        end = min(len(raw) - 1, match.end() + padding)
        text = raw[start:end].replace('\n', ' ').replace('\t', ' ')
        text = text[:terminal_width]
        results.append(highlight_matches(text, cyan))
        
    if len(results) < min_results:
        continue

    results_str = str(len(results))
    if len(results) == max_results:
        results_str += '+'
    print(yellow(f'({results_str}) {path}'))
    print('\n'.join(results[:printed_results]))