{"slug":"schedule-extractor","title":"schedule-extractor","summary":"Extract tabular schedule data from construction drawings — door, window, finish, fixture, panel schedules — and output to Excel. Triggers: 'door schedule', 'extract schedule', 'schedule to Excel', 'panel schedule'.","platform":"Claude","tags":[],"authorName":"LLM Mart","authorSlug":"llm-mart","score":0,"source":"github","price":null,"verified":false,"createdAt":"2026-10-19T21:17:03.319735Z","repo":{"url":"https://github.com/dleerdefi/claude-code-construction","stars":44,"forks":13,"license":"MIT","updatedAt":"2026-10-18T06:12:41Z"},"bodyHtml":"
\n

name: schedule-extractor\ndescription: >\nExtract tabular schedule data from construction drawings — door, window,\nfinish, fixture, panel schedules — and output to Excel. Triggers: 'door\nschedule', 'extract schedule', 'schedule to Excel', 'panel schedule'.\nargument-hint: \"<sheet_number> [schedule_type]\"\ndisable-model-invocation: true

\n

Schedule Extractor

\n

Extracts tabular schedule data embedded in drawing sheets or spec pages and outputs structured Excel files. Schedules are typically one element among many on a sheet — or the entire sheet may be a schedule.

\n

First: the mode (both workflows)

\n

Before extracting or reconciling, read .construction/project.yaml. Present and readable means AgentCM mode: note its contract_version and follow ${CLAUDE_PLUGIN_ROOT}/reference/agentcm-mode.md. Absent means flat-file mode. A listing that shows the file, or .construction/database.yaml alone, does not decide it: read the file.

\n

Workflow

\n
Extraction Progress:\n- [ ] Step 1: Discover schedule locations\n- [ ] Step 2: Isolate (crop) the schedule region\n- [ ] Step 3: Extract structured data (pdfplumber + vision)\n- [ ] Step 4: Validate and clean data\n- [ ] Step 5: Output to Excel\n- [ ] Step 6: Ingest to AgentCM (AgentCM mode)\n- [ ] Step 7: Write a finding\n
\n

Step 1: Discover Schedule Locations

\n

Use a multi-source discovery approach, checking all available sources:

\n

Source A — Sheet titles in sheet index (primary, most reliable)

\n

Search .construction/index/sheet_index.yaml (or the sheet index you've built) for sheets with \"SCHEDULE\" in the title. Many important schedules occupy an entire sheet — the sheet title tells you exactly what it is:

\n\n

For dedicated schedule sheets, the entire page is the extraction target — no need to crop.

\n

Source B — AgentCM's tables (AgentCM mode)

\n

In AgentCM mode (.construction/project.yaml exists; follow ${CLAUDE_PLUGIN_ROOT}/reference/agentcm-mode.md), .construction/INDEX.md's Schedules and legends table lists every table AgentCM found, each in its own .construction/schedules/<sheet>-<n>.tsv. The file's comment line gives the table's box on the sheet (x y w h, 0 to 1 from the top left) and, for a schedule already ingested, its schedule_id (none for a table read from the sheet's text alone). For embedded schedules on non-schedule sheets, that box is where to look: sh .construction/crop.sh <sheet> <x> <y> <w> <h>.

\n

Note: A table read from the text is AgentCM's reading, and its box an estimate: treat both as leads, not definitive boundaries. Always verify with vision.

\n

Source C — Discipline-based heuristics

\n

Schedules appear on specific sheet types:

\n\n

Source D — Vision scan (fallback)

\n

If no index or graph is available:

\n
\"${CLAUDE_PLUGIN_ROOT}/bin/construction-python\" \"${CLAUDE_PLUGIN_ROOT}/scripts/pdf/rasterize_page.py\" \"{pdf_path}\" {page} --dpi 150 --output \".construction/skills/schedule-extractor/pages/full_sheet.png\"\n
\n

Use vision on the full sheet image: \"Identify any tabular schedules on this drawing sheet. Report the approximate bounding box (top-left x,y and bottom-right x,y) as fractions of the image width and height between 0 and 1, the schedule type, and the column headers visible.\"

\n

Step 2: Isolate the Schedule Region

\n

For dedicated schedule sheets (entire page is a schedule): Skip cropping — use the full page.

\n

For embedded schedules (schedule is one element on a larger sheet):

\n

Crop the identified region with padding. The box is the 0–1 fractions from\nStep 1, so --normalized is required; without it the script reads the values\nas pixels and returns an empty crop:

\n
\"${CLAUDE_PLUGIN_ROOT}/bin/construction-python\" \"${CLAUDE_PLUGIN_ROOT}/scripts/pdf/crop_region.py\" \".construction/skills/schedule-extractor/pages/full_sheet.png\" \\\n  --box {x1},{y1},{x2},{y2} --normalized \\\n  --padding 20 \\\n  --output \".construction/skills/schedule-extractor/pages/schedule_crop.png\"\n
\n

Re-rasterize at higher DPI (300) for the cropped region to improve text clarity.\nrasterize_page.py --crop takes percentages, so multiply each fraction by 100\n(0.62,0.08,0.98,0.45 becomes 62,8,98,45):

\n
\"${CLAUDE_PLUGIN_ROOT}/bin/construction-python\" \"${CLAUDE_PLUGIN_ROOT}/scripts/pdf/rasterize_page.py\" \"{pdf_path}\" {page} \\\n  --dpi 300 \\\n  --crop {x1_pct},{y1_pct},{x2_pct},{y2_pct} \\\n  --output \".construction/skills/schedule-extractor/pages/schedule_hires.png\"\n
\n

Step 3: Extract Structured Data

\n

In AgentCM mode — the table AgentCM read

\n

A table INDEX.md lists under Schedules and legends is already read. Turn its file into the exporter's data with the reader, never with a converter of your own:

\n
\"${CLAUDE_PLUGIN_ROOT}/bin/construction-python\" \"${CLAUDE_SKILL_DIR}/scripts/agentcm_schedule.py\" \".construction/schedules/<sheet>-<n>.tsv\" \\\n  --type door_schedule --data \".construction/skills/schedule-extractor/<sheet>-<n>.json\"\n
\n

It prints the header row it took and what it left out. A table read from the sheet's text has its title and header rows at the top: when it took the wrong row, give --header-row <y> (the line's position, as the file prints it) and --header-rows <n> for a header on several lines. Then verify the table against the drawing with crop.sh and the box on the file's comment line (Source B). Where the drawing shows a different value, correct that cell in the data JSON with the Edit tool, and record the correction as ${CLAUDE_PLUGIN_ROOT}/reference/agentcm-mode.md says. Methods A to C below are for a table AgentCM did not find.

\n

Otherwise use a try → validate → fallback approach:

\n

Method A — pdfplumber table extraction (try first)

\n

Extract the page's tables with the shared script (its JSON holds text and tables; add --bbox x1,y1,x2,y2 in PDF points to limit it to the schedule region):

\n
\"${CLAUDE_PLUGIN_ROOT}/bin/construction-python\" \"${CLAUDE_PLUGIN_ROOT}/scripts/pdf/extract_text_region.py\" \"{pdf_path}\" {page} --output \".construction/skills/schedule-extractor/tables.json\"\n
\n

If multiple tables are found, select the largest one (most rows with the most columns — schedules are wide). The first row of the selected table contains headers; subsequent rows are data. If no tables are found or results look garbled, fall back to vision (Method B). Do not write pdfplumber code of your own.

\n

Evaluate Method A — Quality Gate

\n

Check these criteria before proceeding:

\n
    \n
  1. Row count: Did pdfplumber extract at least 10 data rows? Most schedules have 20-200+ entries.
    2. \n
  2. Column consistency: Do ≥80% of rows have the same number of columns? Inconsistent columns = parsing error.
    3. \n
  3. Non-empty cells: Are >50% of cells non-empty? Mostly-empty extraction = parsing failed.
    4. \n
  4. First column validity: Does the first column contain recognizable IDs (door numbers, room numbers, equipment tags)?
    5. \n
\n

Decision:

\n\n
QUALITY GATE:\n  extracted_rows >= 10?          YES → check next  |  NO → use Method B\n  column_consistency >= 80%?     YES → check next  |  NO → use Method B\n  non_empty_cells >= 50%?        YES → check next  |  NO → use Method B\n  first_col_has_valid_ids?       YES → use Method A |  NO → use Method B\n
\n

Method B — Vision extraction (fallback)

\n

When pdfplumber fails (common with complex layouts, merged cells, non-standard table lines):

\n
    \n
  1. Rasterize the sheet at 150 DPI for overview:
    2. \n
\n
\"${CLAUDE_PLUGIN_ROOT}/bin/construction-python\" \"${CLAUDE_PLUGIN_ROOT}/scripts/pdf/rasterize_page.py\" \"{pdf_path}\" {page} --dpi 150 --output \".construction/skills/schedule-extractor/pages/full_sheet.png\"\n
\n
    \n
  1. Use vision on the full sheet image to read the schedule directly:
    2. \n
\n

\"This is a construction drawing sheet containing a schedule (tabular data). Extract ALL rows from the schedule as a JSON array of objects. The first row of the table contains column headers — use those as object keys. Each subsequent row is one entry. Preserve exact values including dimensions, abbreviations, and codes. If a cell is empty, use null.\"

\n
    \n
  1. For dense schedules with small text, re-rasterize at 300 DPI and crop to just the schedule region:
    2. \n
\n
\"${CLAUDE_PLUGIN_ROOT}/bin/construction-python\" \"${CLAUDE_PLUGIN_ROOT}/scripts/pdf/rasterize_page.py\" \"{pdf_path}\" {page} --dpi 300 --output \".construction/skills/schedule-extractor/pages/schedule_hires.png\"\n
\n

Method C — Hybrid (for maximum accuracy)

\n

Run both methods and cross-validate:

\n\n

Step 4: Validate and Clean

\n

After extraction (by either method):

\n\n

Step 5: Output to Excel

\n
\"${CLAUDE_PLUGIN_ROOT}/bin/construction-python\" \"${CLAUDE_SKILL_DIR}/scripts/schedule_to_xlsx.py\" \\\n  --data \".construction/skills/schedule-extractor/<sheet>-<n>.json\" \\\n  --type door_schedule \\\n  --project \"Project Name\" \\\n  --sheet \"A-0.01\" \\\n  --output \"Door_Schedule_A-0.01.xlsx\"\n
\n

The data JSON is working data: write it (with the Write tool, or the reader in AgentCM mode) under .construction/skills/schedule-extractor/, never in the project root or a folder of your own. The workbook is the deliverable: write it in the project folder as the command above shows, never under .construction/. For a table AgentCM has already ingested (schedule_id=<id> on its file's comment line), add --schedule-id <id> --sheet-id <the sheet's id> so the workbook can come back through the round trip.

\n

The script creates a formatted Excel workbook with:

\n\n

Step 6: Ingest to AgentCM (AgentCM mode)

\n

After Excel output, send the extracted data to AgentCM's schedule ingest, which stores it and links its rows to rooms. The payload (below) goes in .construction/skills/schedule-extractor/outbox/schedules-ingest-<sheet>-<YYYYMMDD-HHMMSS>.json:

\n\n

Send it as ${CLAUDE_PLUGIN_ROOT}/reference/agentcm-mode.md says under Writing back; if AgentCM is closed, the payload waits there and the PE is told:

\n
# api_url and project_id from .construction/database.yaml\ncurl -sS -X POST \"{api_url}/projects/{project_id}/schedules/ingest\" \\\n  -H \"Content-Type: application/json\" \\\n  -d @\".construction/skills/schedule-extractor/outbox/schedules-ingest-<sheet>-<YYYYMMDD-HHMMSS>.json\" \\\n  -w \"\\nHTTP %{http_code}\\n\"\n
\n

The sheet's id. sheetId is the id of the entry in .construction/index/sheet_index.yaml whose sheet_number is the schedule's sheet. Never guess it from the number. If that file is missing, or its entries have no id (an older sheet-splitter wrote a pages: list with filename and page_index), say so and stop: do not ingest. The Excel workbook from Step 5 stands either way.

\n

The ingest payload JSON follows this structure:

\n
{\n  \"sheetId\": \"<uuid of the source sheet>\",\n  \"scheduleType\": \"door\",\n  \"title\": \"DOOR SCHEDULE\",\n  \"rowEntityType\": \"door\",\n  \"columns\": [\n    {\"key\": \"MARK\", \"header\": \"MARK\"},\n    {\"key\": \"WIDTH\", \"header\": \"WIDTH\"},\n    {\"key\": \"HEIGHT\", \"header\": \"HEIGHT\"},\n    {\"key\": \"TYPE\", \"header\": \"TYPE\"},\n    {\"key\": \"FRAME\", \"header\": \"FRAME\"},\n    {\"key\": \"HARDWARE_SET\", \"header\": \"HARDWARE SET\"},\n    {\"key\": \"FIRE_RATING\", \"header\": \"FIRE RATING\"}\n  ],\n  \"rows\": [\n    {\n      \"entityIdentifier\": \"2.101A\",\n      \"cells\": {\"MARK\": \"2.101A\", \"WIDTH\": \"3'-0\\\"\", \"HEIGHT\": \"7'-0\\\"\", \"TYPE\": \"A\", \"FRAME\": \"HM\", \"HARDWARE_SET\": \"HW-3\", \"FIRE_RATING\": \"1 HR\"}\n    }\n  ],\n  \"sourceMethod\": \"pdfplumber\",\n  \"confidence\": 0.95\n}\n
\n

Schedule types and their rowEntityType:

\n\n

The ingest endpoint automatically:

\n
    \n
  1. Creates/updates the schedule record
    2. \n
  2. Inserts all rows with GIN-indexed jsonb cells
    3. \n
  3. Links rows to rooms (exact match, fuzzy match, or creates rooms from schedule data)
    4. \n
  4. Extracts finish/hardware codes into the code legends table
    5. \n
  5. Creates pending links for fixtures/equipment that need spatial resolution
    6. \n
\n

Step 7: Write a Finding

\n

In both modes, record the extraction in .construction/skills/schedule-extractor/findings/:

\n
\"${CLAUDE_PLUGIN_ROOT}/bin/construction-python\" \"${CLAUDE_PLUGIN_ROOT}/scripts/graph/write_finding.py\" \\\n  --skill \"schedule-extractor\" --type \"schedule_extracted\" \\\n  --title \"Door schedule extracted from A-0.01\" \\\n  --source-sheet \"A-0.01\" \\\n  --output-file \"Door_Schedule_A-0.01.xlsx\" \\\n  --data '{\"schedule_type\": \"door\", \"row_count\": 45, \"columns\": [\"MARK\",\"SIZE\",\"TYPE\",\"FRAME\",\"HARDWARE SET\"]}'\n
\n

Schedule Reconciliation (Excel → DB Round-Trip)

\n

When a PE edits an exported Excel schedule and wants changes applied back to AgentCM:

\n

Reconciliation Workflow

\n
Reconciliation Progress:\n- [ ] Step R1: Read metadata from edited Excel\n- [ ] Step R2: Compute changeset (diff vs DB)\n- [ ] Step R3: POST changeset to reconcile endpoint\n- [ ] Step R4: Report summary to user\n
\n

Step R1: Read Metadata from Edited Excel

\n

Exported Excel files contain reconciliation anchors:

\n\n

If the Excel lacks _agentcm_meta, it was not exported by AgentCM — treat as a fresh import (re-run the extraction workflow above).

\n

Step R2: Compute Changeset

\n

Run the diff engine to compare the edited Excel against the rows AgentCM holds:

\n
\"${CLAUDE_PLUGIN_ROOT}/bin/construction-python\" \"${CLAUDE_SKILL_DIR}/scripts/xlsx_to_changeset.py\" \\\n  --excel \"edited_schedule.xlsx\" \\\n  --output \".construction/skills/schedule-extractor/outbox/schedules-reconcile-<sheet>-<YYYYMMDD-HHMMSS>.json\"\n
\n

The script:

\n
    \n
  1. Reads _agentcm_meta → extracts schedule_id. A table never ingested (schedule_id=none) has nothing to reconcile and is refused: ingest it first (Step 6)
    2. \n
  2. Reads data sheet → extracts _row_key column + all data columns
    3. \n
  3. Reads the schedule's stored rows from AgentCM, GET {api_url}/projects/{project_id}/schedules/{schedule_id}/rows, with project_id and api_url from .construction/database.yaml (nothing else in it)
    4. \n
  4. Computes diff: cell changes, new rows, deleted rows, new columns, hidden columns
    5. \n
  5. Outputs a structured JSON changeset, with sourceFile the edited Excel's name
    6. \n
\n

It needs AgentCM running. If AgentCM does not answer, the script exits 3 and computes nothing: tell the PE to start AgentCM and rerun this step (the edited Excel is all it needs).

\n

Step R3: POST Changeset to Reconcile Endpoint

\n

Send the changeset R2 wrote to the outbox, as ${CLAUDE_PLUGIN_ROOT}/reference/agentcm-mode.md says under Writing back:

\n
# api_url and project_id from .construction/database.yaml\ncurl -sS -X POST \"{api_url}/projects/{project_id}/schedules/reconcile\" \\\n  -H \"Content-Type: application/json\" \\\n  -d @\".construction/skills/schedule-extractor/outbox/schedules-reconcile-<sheet>-<YYYYMMDD-HHMMSS>.json\" \\\n  -w \"\\nHTTP %{http_code}\\n\"\n
\n

The reconcile endpoint applies changes with:

\n\n

Step R4: Report Summary

\n

The endpoint returns an ExcelReconcileResult:

\n
{\n  \"rowsUpdated\": 3,\n  \"rowsAdded\": 1,\n  \"rowsSoftDeleted\": 0,\n  \"cellsChanged\": 7,\n  \"conflictsCreated\": 0,\n  \"columnsAdded\": 0,\n  \"columnsHidden\": 0,\n  \"errors\": []\n}\n
\n

Report the summary to the user. If errors is non-empty, surface those for review.

\n

Anchored Excel Export

\n

When exporting schedules for PE review, always include reconciliation anchors:

\n
\"${CLAUDE_PLUGIN_ROOT}/bin/construction-python\" \"${CLAUDE_SKILL_DIR}/scripts/schedule_to_xlsx.py\" \\\n  --data schedule_data.json \\\n  --type door_schedule \\\n  --project \"Project Name\" \\\n  --sheet \"A-0.01\" \\\n  --schedule-id \"{schedule_uuid}\" \\\n  --sheet-id \"{sheet_uuid}\" \\\n  --output \"Door_Schedule_A-0.01.xlsx\"\n
\n

The --schedule-id and --sheet-id flags embed metadata in the hidden _agentcm_meta sheet, enabling the round-trip reconciliation workflow.

\n

Schedule Type Reference

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
TypeCommon columnsNotes
DoorMark, Width, Height, Type, Frame, Hardware Set, Fire RatingCross-ref hardware sets with spec 08 71 00
WindowMark, Width, Height, Type, Glazing, Frame MaterialCheck for energy code compliance
Room FinishRoom #, Name, Floor, Base, North/South/East/West Walls, CeilingFinish codes defined in legend
PanelCircuit #, Description, Load (VA), Breaker Size, PhaseVerify total load vs panel capacity
Fixture (Plumbing)Mark, Description, Manufacturer, Model, ConnectionCross-ref with spec Division 22
Equipment (HVAC)Tag, Description, CFM/BTU, Voltage, WeightCross-ref with spec Division 23
\n

File Safety

\n

Never overwrite an existing schedule extraction. The export script uses safe_output_path() which appends _v2, _v3, etc. automatically.

\n
\n

Allowed Scripts

\n

Allowed scripts — exhaustive list. Only execute these scripts during this skill:

\n\n","files":[{"path":"agents/openai.yaml","sizeBytes":268,"isText":true},{"path":"scripts/agentcm_schedule.py","sizeBytes":9675,"isText":true},{"path":"scripts/schedule_to_xlsx.py","sizeBytes":7815,"isText":true},{"path":"scripts/shared.py","sizeBytes":554,"isText":true},{"path":"scripts/xlsx_to_changeset.py","sizeBytes":12863,"isText":true},{"path":"SKILL.md","sizeBytes":21256,"isText":true}],"reviewScore":null,"reviewSummary":null,"trust":{"provenance":"trusted-source-unreviewed","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow.","bodySource":null},"bodyLocked":false,"purchaseUrl":null,"sourceUrl":null,"report":{"provenance":"trusted-source-unreviewed","screen":{"ran":true,"outcome":"clean","suspicious":0,"notes":0,"hiddenCharacters":false},"virusScan":{"engine":"clamav","status":"clean","scannedAt":"2026-10-19T21:17:14.158065Z","sha256":"BE3C1A36D3DB3B57DF8CDFC1E2C49CBC1E07A872FB4270D3062C476F8753C25C","sizeBytes":19169},"review":null,"source":{"repositoryUrl":"https://github.com/dleerdefi/claude-code-construction","path":"skills/schedule-extractor","license":"MIT","commit":"5c46e16c67cdc82ee8bcbb0fdd8ba316d792d4ec","subtreeSha":"2FE08600BB48767D29735E2DB52D92984D3D2A098870E3E8EE20065C286EEF2F","lastSyncedAt":"2026-10-19T21:17:01.926806Z"},"reviewedAt":"2026-10-19T21:17:40.350764Z","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow."},"install":[{"target":"skills-cli","command":"npx skills add https://github.com/dleerdefi/claude-code-construction/tree/main/skills/schedule-extractor"},{"target":"claude-code","command":"claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install dleerdefi-claude-code-construction@llmmart"},{"target":"git","command":"git clone https://github.com/dleerdefi/claude-code-construction.git"}]}