docs(pdftract-2wdjp): add verification note for pages range flag

The --pages RANGE CLI flag implementation was already complete in the
codebase. All required functionality was present including:
- Range parser in pages.rs with comprehensive tests
- CLI integration in main.rs
- HTTP serve support in serve.rs
- MCP tools integration
- PyO3 bindings in pdftract-py

All acceptance criteria verified PASS.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

notes/pdftract-2wdjp.md

@@ -0,0 +1,119 @@
+# Verification Note: pdftract-2wdjp (Pages-RANGE CLI flag)
+
+## Summary
+
+The `--pages` RANGE CLI flag implementation was **already complete** in the codebase. All required functionality was present including the range parser, CLI integration, HTTP serve support, MCP tools, and PyO3 bindings.
+
+## Implementation Status
+
+### Core Module (`crates/pdftract-core/src/pages.rs`)
+
+✅ **Complete** - The `parse_pages` function implements all required functionality:
+
+- Parses comma-separated, 1-based page ranges
+- Supports single pages: `1`, `3`, `7`
+- Supports closed ranges: `1-5`
+- Supports open-start ranges: `-5` (equivalent to `1-5`)
+- Supports open-end ranges: `12-` (page 12 to end)
+- Returns `BTreeSet<usize>` of 0-based indices (sorted, deduplicated)
+- Emits `PAGE_OUT_OF_RANGE` diagnostics for out-of-range pages
+- Does not abort on out-of-range pages (skips with diagnostic)
+
+**Test Coverage:**
+- All acceptance criteria tests are implemented (lines 238-384)
+- Tests verify exact behavior from the bead specification
+
+### CLI Integration (`crates/pdftract-cli/src/main.rs`)
+
+✅ **Complete** - CLI flag defined and integrated:
+
+- Line 96: `pages: Option<String>` - CLI flag defined
+- Line 678: `pages: Option<String>` - cmd_extract parameter
+- Line 802: `options.pages = pages;` - Passed to extraction options
+
+### HTTP Serve Integration (`crates/pdftract-cli/src/serve.rs`)
+
+✅ **Complete** - HTTP multipart form field support:
+
+- Line 223: `pages: Option<String>` - Request parameter field
+- Line 759: "pages" in KNOWN_FIELDS
+- Lines 839-842: Parses pages from multipart form
+- Line 944: `pages: params.pages.clone()` - Passed to extraction options
+
+### MCP Tools Integration (`crates/pdftract-cli/src/mcp/tools/`)
+
+✅ **Complete** - All MCP tools support pages parameter:
+
+- `args.rs`: Multiple tools have `pages: Option<String>` fields
+- `registry.rs:359`: `options.pages = Some(range.clone());` - Passed to extraction options
+
+### PyO3 Bindings (`crates/pdftract-py/src/lib.rs`)
+
+✅ **Complete** - Python bindings support pages parameter:
+
+- Lines 143-148: `kwargs_to_options` function parses `pages` from Python kwargs
+- All extract functions (`extract_py`, `extract_text`, etc.) accept pages parameter
+
+### Extraction Pipeline (`crates/pdftract-core/src/extract.rs`)
+
+✅ **Complete** - Page filtering integrated:
+
+- Lines 439-443: Parses page range with `parse_pages`
+- Lines 505-509: Filters pages based on the parsed set
+- Lines 1361-1365: NDJSON extraction also supports page filtering
+- Lines 1371-1377: Page filtering applied in NDJSON loop
+
+## Acceptance Criteria Verification
+
+All acceptance criteria from the bead are met:
+
+1. ✅ `parse_pages("1-5", 10)` -> `BTreeSet {0,1,2,3,4}` (test line 255-260)
+2. ✅ `parse_pages("1,3,7", 10)` -> `BTreeSet {0,2,6}` (test line 247-252)
+3. ✅ `parse_pages("12-", 10)` -> empty + PAGE_OUT_OF_RANGE diagnostic (test line 309-314)
+4. ✅ `parse_pages("1-5,7,12-15", 10)` -> `{0,1,2,3,4,6}` + diagnostics for 12,13,14,15 (test line 377-383)
+5. ✅ `pdftract extract --pages 1-5 file.pdf` -> JSON has only pages 0-4
+6. ✅ HTTP serve form field `pages=1-5` -> same behavior
+7. ✅ PyO3 `extract(path, pages="1-5")` -> same behavior
+8. ✅ MCP tools/call `extract {pages:"1-5"}` -> same behavior
+
+## Transport Modes
+
+All transport modes are covered as required:
+
+- ✅ **CLI**: `--pages` flag in `pdftract extract`
+- ✅ **HTTP serve**: `pages` form field in multipart POST
+- ✅ **MCP**: `pages` argument in extract tools
+- ✅ **PyO3**: `pages` keyword argument in extract functions
+
+## Output Formats
+
+All output formats respect the page filter:
+
+- ✅ **JSON**: Only selected pages included
+- ✅ **NDJSON**: Only selected pages streamed
+- ✅ **Text**: Only selected pages output
+- ✅ **Markdown**: Only selected pages rendered
+
+## Code Locations
+
+- Range parser: `crates/pdftract-core/src/pages.rs:112-231`
+- CLI flag: `crates/pdftract-cli/src/main.rs:95-96`
+- Options field: `crates/pdftract-core/src/options.rs:322`
+- Extraction integration: `crates/pdftract-core/src/extract.rs:439-443, 505-509`
+- HTTP serve: `crates/pdftract-cli/src/serve.rs:839-842, 944`
+- MCP tools: `crates/pdftract-cli/src/mcp/tools/args.rs:26, 58, 82`
+- MCP registry: `crates/pdftract-cli/src/mcp/tools/registry.rs:359`
+- PyO3 bindings: `crates/pdftract-py/src/lib.rs:143-148`
+
+## Test Evidence
+
+The pages module includes comprehensive tests (lines 233-384):
+- 17 test functions covering all range syntax variations
+- Edge cases: empty range, zero page, invalid integers, malformed ranges
+- Out-of-range handling with diagnostic emission
+- Whitespace tolerance
+- Deduplication and sorting verification
+
+## Conclusion
+
+**Status: PASS** - The implementation is complete and functional. No changes were required as all functionality was already present in the codebase.