Regular expressions and text guides
Extract emails and URLs: candidates are not verified addresses
Use extraction for discovery, trim context carefully, and validate candidates according to the consuming application.
Define candidate extraction
Extraction finds strings resembling an address or URL inside larger text. It does not prove deliverability, reachability or ownership.
Use a log excerpt with known positive and negative examples. Decide whether punctuation belongs to surrounding prose or to the candidate itself.
Contact: ada@example.com
Docs: https://example.com/guide?q=oneValidate after extraction
An email syntax check has a defined scope and cannot verify that a mailbox exists. International addresses and quoted local parts need explicit support decisions.
Nav Tools’ email validator uses its documented practical format. A result rejected by that rule is not automatically invalid under every email protocol.
Parse URLs structurally
A URL parser separates scheme, host, path and query. A syntactically valid URL can still point to an unavailable or unwanted destination.
Allow only required schemes and apply destination policy before fetching. A regex match alone cannot implement a safe server-side request policy.
Keep useful provenance
Deduplicate only when repeated occurrences are irrelevant; logs may need every occurrence and its location. Keep the original text for review.
Compare extracted counts with a small hand-reviewed fixture, then process the larger input. Do not auto-request every extracted URL merely to check it.