feat(config): add [pdf.ocr] section — qwen2.5vl:3b default, opt-in + env overrides + doc(app): PdfOcrOpts field doc (Step 4 I-1)

Step 5 (Group F) of v0.20.0 sub-item 1 (scanned PDF OCR) plan +
Step 4 reviewer Important I-1 fix (PdfOcrOpts field doc) 동봉.

F1 — `kebab-config::PdfCfg` + `PdfOcrCfg` + 4 default fn:
- PdfCfg { ocr: PdfOcrCfg }.
- PdfOcrCfg with 11 field (enabled/always_on/engine/model/endpoint/
  languages/max_pixels/request_timeout_secs/valid_ratio_threshold/
  min_char_count/lang_hint).
- defaults: opt-in (enabled=false), qwen2.5vl:3b, 0.5 threshold, 20 char.
- mirror of image OCR cfg pattern (spec §4.5).

Config struct extension:
- `pdf: PdfCfg` field with `#[serde(default = "PdfCfg::defaults")]`.

11 env var override (parallel to KEBAB_IMAGE_OCR_*):
KEBAB_PDF_OCR_{ENABLED,ALWAYS_ON,ENGINE,MODEL,ENDPOINT,LANGUAGES,
MAX_PIXELS,REQUEST_TIMEOUT_SECS,VALID_RATIO_THRESHOLD,MIN_CHAR_COUNT,
LANG_HINT}.

F2 — `crates/kebab-config/tests/pdf_ocr.rs` (신규):
- toml roundtrip (11 field).
- defaults (opt-in + qwen2.5vl:3b).
- env override (4 key sample + default preservation).

F3 (Step 4 I-1) — `pdf_ocr_apply.rs` 4 public item 의 doc comment:
- PdfOcrOpts struct + 6 field.
- PdfOcrSummary struct + 2 field.
- apply_ocr_to_pdf_pages fn (Errors block 포함).
- PdfOcrProgress enum + 2 variant + 5 field.
body 변경 0, doc-only.

spec:  docs/superpowers/specs/2026-05-27-pdf-scanned-ocr-spec.md (§4.5)
plan:  docs/superpowers/plans/2026-05-27-pdf-scanned-ocr-plan.md (Step 5 F1+F2)
prior: 9f003ef (Step 4) — code reviewer Important I-1 resolution
contract: §9 (additive minor wire bump — Step 7)

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
2026-05-27 07:07:18 +00:00
parent 9f003ef1cd
commit fd918a60ce
3 changed files with 260 additions and 2 deletions

View File

@@ -22,11 +22,26 @@ use lopdf::Document as LopdfDocument;
use time::OffsetDateTime;
use tracing::warn;
/// Per-page OCR knobs threaded through [`apply_ocr_to_pdf_pages`].
/// Mirrors the `[pdf.ocr]` config block (spec §4.5); the facade
/// (`kebab_app::ingest_one_pdf_asset`) fills these from
/// `kebab_config::Config::pdf::ocr` plus runtime flags (CLI / SIGINT).
pub struct PdfOcrOpts {
/// Master switch. `false` short-circuits to
/// `PdfOcrSummary { pages_ocrd: 0, ms_total: 0 }` without lopdf reparse.
pub enabled: bool,
/// `true` → 모든 page OCR (dual-block path, new `Block::Paragraph` push).
/// `false` → text-detect block 의 `min_char_count` 또는
/// `valid_ratio_threshold` 미달인 page 만 OCR (in-place mutate).
pub always_on: bool,
/// 0.0..=1.0. text-detect block 의 `compute_valid_char_ratio` 가
/// 본 임계 미만이면 OCR fallback. Default `0.5`.
pub valid_ratio_threshold: f32,
/// text-detect block 의 char count 가 본 임계 미만이면 OCR fallback.
/// empty page (cover, blank separator) 자동 skip. Default `20`.
pub min_char_count: u32,
/// OCR engine 에 전달할 언어 힌트 (예: `Lang("kor".into())`).
/// `None` → no hint passed to engine.
pub lang_hint: Option<Lang>,
/// Optional per-page cancellation handle. checked at start of each page
/// loop iteration; set→true 시 `cancelled mid-PDF` error 반환. plan §6 E4
@@ -34,12 +49,34 @@ pub struct PdfOcrOpts {
pub cancel: Option<Arc<AtomicBool>>,
}
/// OCR run summary returned by [`apply_ocr_to_pdf_pages`] for the caller's
/// `IngestItem.pdf_ocr_pages` + `pdf_ocr_ms_total` wire fields (§4.6.2).
#[derive(Debug)]
pub struct PdfOcrSummary {
/// Number of pages 가 OCR pipeline 을 실제 통과 (skipped page 제외).
pub pages_ocrd: u32,
/// Cumulative wall-clock duration of successful OCR engine calls (ms).
/// `saturating_add` 사용 — 24-day cumulative 까지 overflow-safe.
pub ms_total: u64,
}
/// Post-extract OCR enrichment for PDF. Walks `canonical.blocks` page-by-page,
/// classifies each page via `text_quality::compute_valid_char_ratio` +
/// `min_char_count`, and either:
/// - skips (vector PDF + sufficient text + `always_on=false`),
/// - mutates the text-detect `Block::Paragraph` in-place with OCR output
/// (scanned/mojibake page), or
/// - pushes a new `Block::Paragraph` with dual ordinal (`always_on=true` +
/// vector page).
///
/// Errors:
/// - cancel handle (`opts.cancel = Some(true)`) → `Err("PDF OCR cancelled mid-PDF at page N")`.
/// - lopdf re-parse failure → `Err(...)`.
/// - per-page OCR engine failure 또는 DCTDecode 부재 → `ProvenanceKind::Warning`
/// event push + `emit_progress(Finished { skipped: true })` + continue
/// (no `Err` propagation).
///
/// See spec §4.1 + §4.4 for the full pipeline.
pub fn apply_ocr_to_pdf_pages<F>(
canonical: &mut CanonicalDocument,
engine: &dyn OcrEngine,
@@ -233,12 +270,26 @@ fn find_paragraph_block_idx(blocks: &[Block], page_num: u32) -> usize {
.expect("PdfTextExtractor emits 1 Block::Paragraph per page (invariant)")
}
/// Per-page OCR progress event 가 caller 의 `emit_progress` closure 호출 시 emit.
/// Step 6 의 ingest_one_pdf_asset 가 IngestEvent::PdfOcrStarted / PdfOcrFinished
/// 로 carry (spec §4.6.1 wire schema).
pub enum PdfOcrProgress {
Started { page: u32 },
Finished {
/// page 별 OCR 시작 시 emit. `engine.recognize` 호출 직전.
Started {
/// 1-based PDF page number.
page: u32,
},
/// page 별 OCR 종료 시 emit (성공 / skip / failure 모두).
Finished {
/// 1-based PDF page number.
page: u32,
/// `engine.recognize` wall-clock duration. skip path 의 의미는 mixed
/// (DCTDecode 부재 시 `0`, OCR engine 실패 시 actual latency before bail).
ms: u64,
/// OCR result text 의 char count. skip 시 `0`.
chars: u32,
/// `true` = DCTDecode 부재 또는 OCR engine 실패 로 skip.
/// `false` = 정상 OCR 완료.
skipped: bool,
},
}

View File

@@ -50,6 +50,11 @@ pub struct Config {
/// load cleanly with built-in defaults.
#[serde(default)]
pub ingest: IngestCfg,
/// v0.20.0 sub-item 1: PDF ingest pipeline settings. `#[serde(default)]`
/// so pre-v0.20 config files without a `[pdf]` section load with
/// built-in defaults (OCR disabled — opt-in for scanned PDF KB).
#[serde(default = "PdfCfg::defaults")]
pub pdf: PdfCfg,
/// p9-fb-05: directory of the on-disk config file this `Config`
/// was loaded from, if any. Populated by `Config::from_file` /
/// `Config::load` — never serialized (`#[serde(skip)]`). Used by
@@ -392,6 +397,88 @@ impl CaptionCfg {
}
}
/// Settings for the PDF ingest pipeline (P7 + v0.20.0 sub-item 1).
#[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
pub struct PdfCfg {
#[serde(default = "PdfOcrCfg::defaults")]
pub ocr: PdfOcrCfg,
}
impl PdfCfg {
pub fn defaults() -> Self {
Self { ocr: PdfOcrCfg::defaults() }
}
}
impl Default for PdfCfg {
fn default() -> Self { Self::defaults() }
}
/// v0.20.0 sub-item 1: scanned PDF OCR via Ollama vision LLM. Default
/// disabled — opt-in because OCR adds ~45-100s per scanned page on CPU
/// (qwen2.5vl:3b, remote). Enable for book / paper scan KB.
#[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
pub struct PdfOcrCfg {
/// Run OCR on scanned PDF pages. Default `false` (opt-in).
pub enabled: bool,
/// `false` (default) — text-detect first + vision fallback on
/// scanned pages only. `true` — vision LLM 호출 on every page
/// (vector PDF 의 dual-text confidence boost — doubles chunk count).
pub always_on: bool,
/// Engine identifier. v1 only ships `"ollama-vision"`.
pub engine: String,
/// Vision model id. Default `"qwen2.5vl:3b"` per PoC (§3.5 family
/// asymmetry vs image OCR's gemma4:e4b is acknowledged).
pub model: String,
/// HTTP endpoint. `None` → fall back to `models.llm.endpoint`.
#[serde(default)]
pub endpoint: Option<String>,
/// BCP-47 language hints rendered into prompt.
pub languages: Vec<String>,
/// Long-edge cap (px). Larger images bloat prompt cost.
pub max_pixels: u32,
/// HTTP request timeout (sec). Same `0` = "fail immediately"
/// semantics as `image.ocr.request_timeout_secs` (NOT a disable
/// sentinel — see image.ocr docs).
#[serde(default = "default_pdf_ocr_request_timeout_secs")]
pub request_timeout_secs: u64,
/// Valid char ratio threshold (0.0..=1.0). Page with ratio below
/// this is classified as scanned/mojibake → OCR fallback. Default
/// `0.5`.
#[serde(default = "default_pdf_ocr_valid_ratio")]
pub valid_ratio_threshold: f32,
/// Minimum char count per page below which page is auto-scanned.
/// Default `20`.
#[serde(default = "default_pdf_ocr_min_char_count")]
pub min_char_count: u32,
/// Single-page lang hint. Default `Some("kor")`. `None` = no hint.
#[serde(default = "default_pdf_ocr_lang_hint")]
pub lang_hint: Option<String>,
}
impl PdfOcrCfg {
pub fn defaults() -> Self {
Self {
enabled: false,
always_on: false,
engine: "ollama-vision".to_string(),
model: "qwen2.5vl:3b".to_string(),
endpoint: None,
languages: vec!["eng".to_string(), "kor".to_string()],
max_pixels: 2048,
request_timeout_secs: default_pdf_ocr_request_timeout_secs(),
valid_ratio_threshold: default_pdf_ocr_valid_ratio(),
min_char_count: default_pdf_ocr_min_char_count(),
lang_hint: default_pdf_ocr_lang_hint(),
}
}
}
fn default_pdf_ocr_request_timeout_secs() -> u64 { 600 }
fn default_pdf_ocr_valid_ratio() -> f32 { 0.5 }
fn default_pdf_ocr_min_char_count() -> u32 { 20 }
fn default_pdf_ocr_lang_hint() -> Option<String> { Some("kor".to_string()) }
/// p9-fb-14: TUI-only configuration. Currently a single `theme`
/// selector (`"dark"` / `"light"`); future fields (custom role
/// overrides, mode-machine cursor shapes, …) extend the same
@@ -539,6 +626,7 @@ impl Config {
image: ImageCfg::defaults(),
ui: UiCfg::defaults(),
ingest: IngestCfg::default(),
pdf: PdfCfg::defaults(),
// p9-fb-05: defaults are not loaded from disk, so no
// source_dir. Relative `workspace.root` (rare with
// defaults) falls back to caller `cwd` via the
@@ -903,6 +991,45 @@ impl Config {
self.image.caption.prompt_template_version = v.clone();
}
// pdf.ocr (v0.20.0 sub-item 1)
"KEBAB_PDF_OCR_ENABLED" => self.pdf.ocr.enabled = parse_bool(v),
"KEBAB_PDF_OCR_ALWAYS_ON" => self.pdf.ocr.always_on = parse_bool(v),
"KEBAB_PDF_OCR_ENGINE" => self.pdf.ocr.engine = v.clone(),
"KEBAB_PDF_OCR_MODEL" => self.pdf.ocr.model = v.clone(),
"KEBAB_PDF_OCR_ENDPOINT" => {
self.pdf.ocr.endpoint = if v.is_empty() { None } else { Some(v.clone()) };
}
"KEBAB_PDF_OCR_LANGUAGES" => {
self.pdf.ocr.languages = v
.split(',')
.map(|s| s.trim().to_string())
.filter(|s| !s.is_empty())
.collect();
}
"KEBAB_PDF_OCR_MAX_PIXELS" => {
if let Ok(n) = v.parse::<u32>() {
self.pdf.ocr.max_pixels = n;
}
}
"KEBAB_PDF_OCR_REQUEST_TIMEOUT_SECS" => {
if let Ok(n) = v.parse::<u64>() {
self.pdf.ocr.request_timeout_secs = n;
}
}
"KEBAB_PDF_OCR_VALID_RATIO_THRESHOLD" => {
if let Ok(n) = v.parse::<f32>() {
self.pdf.ocr.valid_ratio_threshold = n.clamp(0.0, 1.0);
}
}
"KEBAB_PDF_OCR_MIN_CHAR_COUNT" => {
if let Ok(n) = v.parse::<u32>() {
self.pdf.ocr.min_char_count = n;
}
}
"KEBAB_PDF_OCR_LANG_HINT" => {
self.pdf.ocr.lang_hint = if v.is_empty() { None } else { Some(v.clone()) };
}
// Unknown KEBAB_* keys are silently ignored — see
// `env_unknown_key_is_ignored` test.
_ => {}

View File

@@ -0,0 +1,80 @@
// crates/kebab-config/tests/pdf_ocr.rs
//
// Integration tests for [pdf.ocr] config section (v0.20.0 sub-item 1).
use std::collections::HashMap;
use kebab_config::{Config, PdfCfg};
// Test 1: toml roundtrip — spec §4.5 line 1034-1047 example block.
// Config requires many required fields; test the [pdf] section via PdfCfg wrapper.
#[derive(serde::Deserialize)]
struct PdfWrapper { pdf: PdfCfg }
#[test]
fn pdf_ocr_toml_roundtrip() {
let toml = r#"
[pdf.ocr]
enabled = true
always_on = false
engine = "ollama-vision"
model = "qwen2.5vl:7b"
endpoint = "http://192.168.0.47:11434"
languages = ["eng", "kor"]
max_pixels = 3072
request_timeout_secs = 900
valid_ratio_threshold = 0.6
min_char_count = 30
lang_hint = "kor"
"#;
let w: PdfWrapper = toml::from_str(toml).expect("parse toml");
let ocr = &w.pdf.ocr;
assert!(ocr.enabled);
assert!(!ocr.always_on);
assert_eq!(ocr.engine, "ollama-vision");
assert_eq!(ocr.model, "qwen2.5vl:7b");
assert_eq!(ocr.endpoint.as_deref(), Some("http://192.168.0.47:11434"));
assert_eq!(ocr.languages, vec!["eng".to_string(), "kor".to_string()]);
assert_eq!(ocr.max_pixels, 3072);
assert_eq!(ocr.request_timeout_secs, 900);
assert!((ocr.valid_ratio_threshold - 0.6).abs() < 1e-6);
assert_eq!(ocr.min_char_count, 30);
assert_eq!(ocr.lang_hint.as_deref(), Some("kor"));
}
// Test 2: defaults — opt-in, qwen2.5vl:3b model, 0.5 threshold, 20 min_char.
#[test]
fn pdf_ocr_defaults_off_with_qwen_3b() {
let cfg = Config::defaults();
assert_eq!(cfg.pdf.ocr.enabled, false);
assert_eq!(cfg.pdf.ocr.always_on, false);
assert_eq!(cfg.pdf.ocr.engine, "ollama-vision");
assert_eq!(cfg.pdf.ocr.model, "qwen2.5vl:3b");
assert!(cfg.pdf.ocr.endpoint.is_none());
assert_eq!(cfg.pdf.ocr.languages, vec!["eng".to_string(), "kor".to_string()]);
assert_eq!(cfg.pdf.ocr.max_pixels, 2048);
assert_eq!(cfg.pdf.ocr.request_timeout_secs, 600);
assert!((cfg.pdf.ocr.valid_ratio_threshold - 0.5).abs() < 1e-6);
assert_eq!(cfg.pdf.ocr.min_char_count, 20);
assert_eq!(cfg.pdf.ocr.lang_hint.as_deref(), Some("kor"));
}
// Test 3: env var override — 4 keys 의 typical override case.
#[test]
fn pdf_ocr_env_overrides() {
let mut env: HashMap<String, String> = HashMap::new();
env.insert("KEBAB_PDF_OCR_ENABLED".to_string(), "true".to_string());
env.insert("KEBAB_PDF_OCR_MODEL".to_string(), "qwen2.5vl:7b".to_string());
env.insert("KEBAB_PDF_OCR_ALWAYS_ON".to_string(), "true".to_string());
env.insert("KEBAB_PDF_OCR_VALID_RATIO_THRESHOLD".to_string(), "0.75".to_string());
let cfg = Config::defaults().apply_env(&env);
assert_eq!(cfg.pdf.ocr.enabled, true);
assert_eq!(cfg.pdf.ocr.model, "qwen2.5vl:7b");
assert_eq!(cfg.pdf.ocr.always_on, true);
assert!((cfg.pdf.ocr.valid_ratio_threshold - 0.75).abs() < 1e-6);
// 다른 env var 가 default 보존
assert_eq!(cfg.pdf.ocr.engine, "ollama-vision");
assert_eq!(cfg.pdf.ocr.min_char_count, 20);
}