docs: add Repository Guide to README describing all files and folders#149
docs: add Repository Guide to README describing all files and folders#149Copilot wants to merge 2 commits into
Conversation
…es and folders Agent-Logs-Url: https://github.com/RConsortium/submissions-pilot5-datasetjson/sessions/63ccb6be-e02f-4141-b75a-f8674b710526 Co-authored-by: bms63 <10111024+bms63@users.noreply.github.com>
Dataset QC Check Complete!Date: 2026-05-05 ✅ Dataset: adadas✅ Dataset: adae✅ Dataset: adlbc✅ Dataset: adsl✅ Dataset: adtte |
TLF QC Check Complete!Date: 2026-05-05 ✅ Dataset: tlf-demographic-pilot5.out✅ TLF: tlf-kmplot-pilot5.pdf - Score: 85[1] "1. Title differences: First image shows 'KM plot for Time to First Dermatologic Event: Safety population' while second shows 'KM plot for Time to First Dermatologic Event-Safety population' (hyphen vs colon)\n\n2. Grid lines: The second image has a visible grid overlay, while the first image has a cleaner background without prominent grid lines\n\n3. Legend positioning: The legend appears in different positions - first image has legend on the right side, second image also has legend on right but positioned slightly differently\n\n4. Confidence intervals: Both show shaded confidence intervals but the visual rendering appears slightly different in opacity/shading\n\n5. Axis labels and formatting: Minor differences in axis label positioning and formatting\n\n6. Line styling: The curves appear to have slightly different line weights or rendering\n\n7. Table formatting below chart: The 'At risk' tables below the charts show the same data but with different formatting and spacing\n\n8. Y-axis label: First image shows 'Probability of event' while formatting may differ slightly\n\n9. Program information at bottom: Different program identifiers and timestamps - first shows 'Program ID: tmplot.r [ 2024-01-07 18:07:04]662' while second shows 'Program: tfl~kmplot.r [ 2024-04-18 20:17:28]'\n\n10. Overall styling: Second image appears to have a more prominent grid structure making it look more 'technical' while first has cleaner aesthetic" ✅ TLF: tlf-efficacy-pilot5.rtf - Score: 99[1] "After careful examination of both images, I found the following differences:\n\n1. Timestamp/Date difference: The bottom of the table contains different program execution timestamps. The first image shows '...2024-04-18 20:17...' while the second image shows '...2024-04-18 20:17...' - upon closer inspection, these appear identical or nearly identical.\n\n2. Image quality/compression: There may be very minor differences in image compression or rendering quality, though both images appear to be of similar quality.\n\n3. The tables themselves appear to be identical in terms of:\n - Title: 'ANCOVA of Change from Baseline at Week 20'\n - Column headers (Week 20, Week 52, Change from Baseline)\n - Treatment rows (Xanomeline High Dose vs Placebo, Xanomeline Low Dose vs Placebo, Xanomeline High Dose vs Median)\n - All numerical values (N, Mean (SE), 95% CI values)\n - Statistical information and footnotes\n - Table structure and formatting\n\nUpon very close inspection, these images appear to be essentially identical with no meaningful differences in the actual content, data, or presentation of the ANCOVA analysis table." ✅ TLF: tlf-primary-pilot5.rtf - Score: 99[1] "After careful examination of both images, I found the following differences:\n\n1. Text in footnote area: In the footnote text at the bottom of the first table, there appears to be a slight difference in one word. The first image contains text that reads '...after a continuous period' while the second image has identical or near-identical text in that section.\n\n2. The footnote text continues with 'CI' notation - both images appear to have 'CI' defined but there may be subtle spacing or formatting differences.\n\n3. Upon very close inspection, the text rendering and anti-aliasing may show minimal pixel-level differences, but this could be due to image compression or capture artifacts rather than actual content differences.\n\nNote: The two images appear to be nearly identical. All numerical values, table headers, column alignments, and structural elements match exactly. The tables show:\n- Protocol CD320020/Page 1 of 2 and Page 2 of 2\n- Population: ITT-E\n- Table 14.2.2\n- Primary Endpoint Analysis: ADAS-Cog - Change from Baseline at Week 24 - LOCF\n- Identical data values across all rows and columns\n- Same footnote structure and content\n\nThe images are essentially duplicates with potentially only minor rendering differences." |
The repo has accumulated many files across ADaM programs, LLM tooling, QC bots, and CI workflows with no single reference explaining what each does or why it exists.
Changes
Adds a Repository Guide section to
README.mdthat catalogues every file and folder grouped by purpose, with a concise description for each entry:.Rprofile,renv.lock,default.nix,build_nixconfig.R,.lintr, etc.original-sdtmdata/,original-adamdata/(read-only Pilot 1/3 reference copies)pilot5-programs/) — fivead*.rscripts (adsl,adae,adlbc,adtte,adadas), each described with its source domains and outputtlf-*.rscripts with their specific output table/figure numbers and packages usedconvert_xpt_to_rds.r,convert_rds_to_json.r,convert_rds_to_xpt.r,convert_rds_to_json_to_xpt.r,convert_xpt_to_datasetjson.r, pluspilot5-helper-fcns.randrun-all-adams-tlfs.rqcReport.qmd(dataset diff viadiffdf) andtlf-qc.qmd(text diff + LLM visual scoring viaellmer/Claude)_adrg.qmd(content),adrg-quarto-html.qmd/adrg-quarto-pdf.qmd(renderers),_publish.ymlllm-adrg-utils/) —llm_pipeline.qmd,llm_api.R,llm_prompts.R,util.R,logging.Rand their CSV outputs inllm-adrg-out/cover-letter/,ectd_readme/README.qmdcreate-ectd-bundle.sh,push-ectd-bundle.sh,create-cmb-report.py) with trigger conditions and purposelogs/llm_calls_<date>.logaudit trail from the LLM pipeline