Creating new publicsuffix file with filled suffix_list_urls

[AnnaSummer]

Hello! I create object of TLDExtract:

extractor = tldextract.TLDExtract(
suffix_list_urls=["file://" + "/absolute/path/to/json/with/custom/public/suffixes"],
cache_dir='/absolute/path/to/cache/dir',
fallback_to_snapshot=False
)

When I call print(extractor("google.ac")), new empty (!) files (lock and json) with public suffixes are created (identifying by new hash in name) and this file is used as a new public suffixes file.

In source code in cache.py (line 108):

cache_filepath = self._key_to_cachefile_path(namespace, key)

method _key_to_cachefile_path create new hash, that is used as a filename of new public suffix file.

Is it a correct behaviour? I just need to avoid HTTP requests to update cache of tlds (in any time including first run).

[john-kurkowski]

When I call print(extractor("google.ac")), new empty (!) files (lock and json) with public suffixes are created

Hello! Are you asking if creating empty cache files is expected behavior? I would say no, that's not expected. Only the files ending in *.lock should be empty, 0 bytes.

Even if this library caches 0 public suffixes, the cache file should be a few bytes, containing the JSON [[], []]. Then calling extractor("google.ac") would raise ValueError: No tlds set. Cannot proceed without tlds..

[john-kurkowski]

So I'm not sure what's going on in your case. Aside, if your file is local anyway, you could play with avoiding the cache step. The cache won't save a ton of processing for a local file.

tldextract.TLDExtract(
    suffix_list_urls=["file://" + "/absolute/path/to/json/with/custom/public/suffixes"],
    cache_dir=None,
    fallback_to_snapshot=False
)

[AnnaSummer]

Thank you for answer!
I need to avoid tlds loading from the Internet (including at first run of library methods), just from local file (only).
I tried to call TLDExtract with your parameters, but get "ValueError: No tlds set. Cannot proceed without tlds."

My local file with public suffixes is not empty (there are tlds in that json).

[john-kurkowski]

there are tlds in that json

That could be the problem. The suffix_list_urls expects to open URLs with plaintext files formatted like the Public Suffix List. Notice that linked file is not JSON.