Geler

Help convert dynamic websites to static ones.

Install

=> dans un venv

pip install geler-CERTIC

=> de manière isolée et indépendante (mais disponible globalement)

pipx install geler-CERTIC

Usage

As a library in your own program:

import geler
freezer = geler.Freezer(
    "https://acme.tld/",
    "/save/to/path/",
    thread_pool_size=10,
    http_get_timeout=30,
    skip_extensions=[".mp4", ".mp3"],
).freeze()

for err in result.http_errors:
    print(f'{err.get("status_code")}: {err.get("url")}')

As a CLI tool:

$> geler --help
usage: geler [-h] [-t THREAD_POOL_SIZE] [--http-get-timeout HTTP_GET_TIMEOUT] [-s SKIP_EXTENSIONS] [-v] start-from-url save-to-path

positional arguments:
  start-from-url        -
  save-to-path          -

optional arguments:
  -h, --help            show this help message and exit
  -t THREAD_POOL_SIZE, --thread-pool-size THREAD_POOL_SIZE
                        1
  --http-get-timeout HTTP_GET_TIMEOUT
                        30
  -s SKIP_EXTENSIONS, --skip-extensions SKIP_EXTENSIONS
                        -
  -v, --verbose         False

Thread pool size (--thread-pool-size) defaults to 1. Increase the number to have multiple downloads in parallel.

HTTP get timeout (--http-get-timeout) default to 30s. This includes the time needed to download the file. Increase the number to increase the timeout, or set it to 0 for no timeout.

List of skipped (--skip-extensions) is a comma-separated list of extensions that won't be downloaded.

Verbose mode (--verbose) will show downloaded URLs and HTTP errors.

Every non-dry run stores its crawl progress in .geler-state.sqlite3 inside the destination directory. If a crawl is interrupted, resume it with the same URL, destination, and skipped extensions:

geler --resume https://acme.tld/ /path/to/local/dir

Resume retries interrupted and failed downloads, skips files that were completed, and patches links again when the crawl finishes. A run without --resume starts a new checkpoint. Geler refuses to resume if the root URL or skipped extensions do not match the checkpoint. Dry runs do not create checkpoints and cannot be resumed.

Complete example:

geler --http-get-timeout 30 --thread-pool-size 10 --skip-extension ".mp4,.zip" https://acme.tld/ /path/to/local/dir

Why ?

For MaX and associated tools, we needed a lightweight, portable, pure Python solution to convert small dynamic websites to static ones.

Alternatives

This tool has a narrow scope, on purpose. Please turn to these solutions if you need more:

• wget
• pywebcopy
• HTTrack

Known Limitations

• only works with HTTP GET
• does not submit forms (even with GET method)
• only considers URLs in src, href, or URL-like JavaScript string literals in onclick attributes
• only considers URLs with http or https schemes
• only downloads what is in the same netloc (same domain, same port) as the start URL
• patches URLs in *.html, *.css and JavaScript module imports in *.js files, but does not parse arbitrary JS URL construction patterns
• does not throttle requests
• does not respect robots.txt or <meta name="robots" content="noindex"> or any other robot directives.