Fix Trufflehog SARIF Results Bugs

[ByteMastermind]

A bug was found when running the trufflehog inside GSAST without parameter --only-verified. After further investigation, several more bugs were found, namely:

• Rule ID mapper key mismatch - The detectors_count_mapper in convert_trufflehog_to_sarif() was stored using detector_name as key but looked up using rule_id_hash.hexdigest(). Since defaultdict(int) returns 0 for any unseen key, almost all findings ended up with the same rule ID "trufflehog - git 1", causing different detector types to collide and get mixed together.

• Duplicate rules created for every finding - A new rule was appended to driver_rules for every single finding, instead of once per unique detector type. With --only-verified off this meant thousands of duplicate rule entries in the SARIF output.

• Results storage overwrite - In store_scan_results(), the loop over split SARIF files always used the same key stored_results[scanner_type], so each iteration overwrote the previous one. Only the last split file was kept, the rest were silently lost.

• Empty line handling in has_findings_trufflehog() - The function didn't strip empty lines before calling json.loads(), so a trailing newline in the TruffleHog output caused a JSONDecodeError and the function returned False even when findings existed.

All these bugs should be handled by the suggested changes. I tested ONLY the json -> SARIF "pipeline" of GSAST, did not try it as a whole.