Fix insecure regex in URL validation and normalize repository data

[Uday111-ai]

This PR addresses two issues:

Security/Bug Fix: The regex used to parse GitHub URLs in
gfi/populate.py
was flawed. It used a character class [...] instead of a group
(...)
for the protocol, and lacked a start/end anchor. This allowed invalid URLs (e.g., notgithub.com) to pass validation.

Fix: Updated GH_URL_PATTERN to r"^(?:https?://)?github.com/(?P[\w.-]+)/(?P[\w.-]+)/?$"
Data Consistency: The
data/repositories.toml
file had an inconsistent entry with a protocol prefix.

Fix: Removed https:// from the entry for pyupio/safety to match the project convention.

Verification:

Validated the new regex against various test cases (valid and invalid URLs).
Verified that
gfi/test_data.py
passes with the changes.

[vercel]

@Uday111-ai is attempting to deploy a commit to the DeepSource Team on Vercel.

A member of the Team first needs to authorize it.