(example="jd salinger")
| 10 | |
| 11 | |
| 12 | def test_summarize_document(example="jd salinger"): |
| 13 | |
| 14 | # pull a sample document (or substitute a file_path and file_name of your own) |
| 15 | sample_files_path = Setup().load_sample_files(over_write=False) |
| 16 | |
| 17 | topic = None |
| 18 | query = None |
| 19 | fp = None |
| 20 | fn = None |
| 21 | |
| 22 | if example not in ["jd salinger", "employment terms", "just the comp", "un resolutions"]: |
| 23 | print ("not found example") |
| 24 | return [] |
| 25 | |
| 26 | if example == "jd salinger": |
| 27 | fp = os.path.join(sample_files_path, "SmallLibrary") |
| 28 | fn = "Jd-Salinger-Biography.docx" |
| 29 | topic = "jd salinger" |
| 30 | query = None |
| 31 | |
| 32 | if example == "employment terms": |
| 33 | fp = os.path.join(sample_files_path, "Agreements") |
| 34 | fn = "Athena EXECUTIVE EMPLOYMENT AGREEMENT.pdf" |
| 35 | topic = "executive compensation terms" |
| 36 | query = None |
| 37 | |
| 38 | if example == "just the comp": |
| 39 | fp = os.path.join(sample_files_path, "Agreements") |
| 40 | fn = "Athena EXECUTIVE EMPLOYMENT AGREEMENT.pdf" |
| 41 | topic = "executive compensation terms" |
| 42 | query = "base salary" |
| 43 | |
| 44 | if example == "un resolutions": |
| 45 | fp = os.path.join(sample_files_path, "SmallLibrary") |
| 46 | fn = "N2126108.pdf" |
| 47 | # fn = "N2137825.pdf" |
| 48 | topic = "key points" |
| 49 | query = None |
| 50 | |
| 51 | # optional parameters: 'query' - will select among blocks with the query term |
| 52 | # 'topic' - will pass a topic/issue as the parameter to the model to 'focus' the summary |
| 53 | # 'max_batch_cap' - caps the number of batches sent to the model |
| 54 | # 'text_only' - returns just the summary text aggregated |
| 55 | |
| 56 | kp = Prompt().summarize_document_fc(fp, fn, topic=topic, query=query, text_only=True, max_batch_cap=15) |
| 57 | |
| 58 | print(f"\nDocument summary completed - {len(kp)} Points") |
| 59 | for i, points in enumerate(kp): |
| 60 | print(i, points) |
| 61 | |
| 62 | return 0 |
| 63 | |
| 64 | |
| 65 | if __name__ == "__main__": |
no test coverage detected