o Ai@sddlZddlZddlmZmZddlZddlmZddl m Z m Z m Z m Z mZmZmZmZmZmZddlmZee dZdedejfd d Zd ed eedefd dZdedeeefdefddZdededeefddZdefddZedkreZ e!de dSdS)N)ListDict)OpenAI) OPENAI_API_KEY EMBED_MODEL LLM_MODEL INDEX_DIR ARTIFACTS_DIR CATEGORIESTOP_K_PER_CATEGORYMAX_FINDINGS_PER_CATEGORYRUN_ID INPUT_DIR) FaissStore)api_keyqueryreturncCs*tjjt|gd}tj|jdjgddS)z.Convert a text query into an embedding vector.)modelinputrfloat32)dtype)client embeddingscreaternparraydata embedding)rresprs z1extract_findings_per_category..rG) rloadrr!searchr r5ospathr6r rKreplacemakedirsopenjsondumpdumpsrr7r8rrloadsr9r:r0JSONDecodeErrorr<printr4r )rNrstoreqvechitsbundlehr@entry bundle_pathfrCrresultrEr]valid_evidencerr`r extract_findings_per_categoryush   #    ryc Csg}d}tdtD]5\}}td|dt||}|D]}d|d|d<||d<|d7}||qtd t|d q |D])}d d |d D}t||d<dddd|ddd}|d|d|d<qD|jdddddd|ddD}tt t|||d} t j t ddt j t d} t| d d!d"}tj| |dd#d$Wdn1swYtd%t|td&d'|td(| d)| S)*z Generate the master findings.json file by: 1. Extracting findings for each category 2. Scoring and ranking findings 3. Identifying top 10 critical findings u* 🔍 Extracting findings per category... z Processing: z...zF-03didrNu → Found z findingscSsh|]}|dqS)r%rr^rrr z)generate_findings_json..rGrecurrence_countr[)HighMediumLowseverityrscorecSs|dS)Nrr)xrrr sz(generate_findings_json..T)keyreversecSsg|]}|dqS)r|r)r_rvrrr rar~z*generate_findings_json..N )run_id input_foldertotal_findingsr] top_10_idsrUz findings.jsonrWrXrYF)r\ ensure_asciiu ✅ Total findings: u✅ Top 10 critical findings: z, u✅ findings.json saved at:  )rnr itemsryr5lenr4sortr rrdrgr rer6rhrirj) all_findingsfidrNrr]rvdoc_set sev_scorer findings_dataout_pathrrr generate_findings_jsonsD   r__main__u✅ findings.json generated at )"rirdtypingrrnumpyropenair rag.configrrrrr r r r r rrag.vectorstorerrr=ndarrayr!rFboolrMryr__name__rernrrrr s" 0  1`5