FacereDataset

Facere/FacereDataset

Fork 0

Commit Graph

Author	SHA1	Message	Date
Knowit	eee1a9b97e	crawler: --skip-ext + --max-source-mb gates for batch-50 expansion Two CLI gates needed before scaling Pro batch beyond top-5: --skip-ext mp4,qt,mov (attachment filter) Skips video extensions in attachment download. Phase 1 measurements showed mp4+qt occupy ~54% of attachment storage. Entry still recorded in metadata.json with skipped:ext:<token> so we can re-fetch later if the policy changes. Honors both server-declared `ext` and filename suffix, case-insensitively. --max-source-mb N (Pro source size cap) Trips inside the chain replay loop on encrypted-blob total. On trip: raise ProjectOversizeError, wipe partial source/, append a row to data/state/oshwhub_pro_oversize.jsonl. Lets us shortlist 50+ Pro projects without one X86-board-class outlier (~500 MB) blowing the LFS budget. Std and Pro 2.x legacy are not capped (both <2 MB in sample). Verified: - cap=0 trips on first blob (1.2 MB), source/ wiped, state recorded - cap=100 runs full ESP-VoCat (7.5 MB plain, 278 docs) - skip-ext microtest: 8/8 cases (case-insensitive, declared/suffix fallback, empty-token edge cases) Plan + frozen candidate list for the next 50 projects: - docs/plans/oshwhub_batch50.md - data/state/oshwhub_batch50_candidates.jsonl (gitignore exception added) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>	2026-04-29 00:24:55 +08:00
Knowit	1e06ba6582	crawler: split sleep policy by host — chain blob fetches drop 5s -> 0.2s The Pro modern fetch_pro_modern walks a per-history blob loop on modules.lceda.cn (CDN-flavored host serving AES-encrypted EPRO2 streams). We were sleeping 5s between every blob — same rate we use for the rate-sensitive pro.lceda.cn API host. HAR analysis (proexportNew2.har) shows the editor fires these blobs back-to-back without throttling, so 0.2s is plenty. Walltime drops linearly with chain length: ESP-VoCat (chain=12): 80s sleep -> 22s sleep (-72%) 220V power (chain=28): 160s sleep -> 26s sleep (-84%) X86 board (chain~700, projection): ~1h -> ~3min Verified by re-fetching ESP-VoCat + 220V power: byte-identical output across all per-doc .epro2 files (sha256 match), only fetched_at timestamp differs in manifest.json. Two manifest files re-stamped as proof of the validation runs. API host sleeps (4x 5s in modern fetcher, 7x 5s in legacy fetcher) are unchanged — those go to pro.lceda.cn /api/ which still wants polite QPS<=0.2. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>	2026-04-29 00:09:19 +08:00
Knowit	3282a028c4	Add EasyEDA Pro EPRO2 source ingestion (3/5 batch test) 打通 oshwhub origin=pro 现代 Pro 3.x 工程的 EPRO2 源抓取链路。3/5 modern Pro 项目完整解出（共 8423 docs / 542 MB plain）： - X86 主板 7374 docs / 481 MB plain (chain=85, editor=3.2.15) - 220V 桌面电源 771 docs / 26 MB plain (chain=28, editor=3.2.69) - ESP-VoCat 278 docs / 7.5 MB plain (chain=12, editor=3.2.91) 剩余 2/5 是 legacy Pro 2.x（立创泰山派 RK3566、梁山派），项目 meta 返回 branch_uuid=null + editorVersion="2.1.40"，没有 git-style chain 模型，文档直接挂在 boards[].sch/pcb 字段上，访问端点暂未挖通；元数据落库 metadata.json，source/ 留空。实现要点： - fetch_pro_source(): 4 步流程（project → branch HEAD → structures → /branches/<B>/histories/<HEAD> 即返完整 chain，无需 ?limit 批量端点）+ 逐 history 走 AES-128-GCM 解密（16 字节 IV，pycryptodome 原生支持）+ gunzip + 按 DOCHEAD 切 per-doc EPRO2 流 - EPRO2 解析坑：行末单 `\|` 是行终止符不是字段分隔符，必须先 rstrip("\|") 再 split("\|\|")，否则 payload JSON 解析失败 silently swallow 导致 cur_doc 不设 → 第一轮 X86 板 7374 docs 抽出来只剩 2 个 - docType 实测远不止 BOARD/PCB/SCH/SCH_PAGE，还含 SYMBOL / FOOTPRINT / DEVICE / BLOB / FONT / CONFIG —— Pro 把组件库快照也随项目存到 history，下游做 EPRO2→KiCad 转换时必须先把这些 lib doc 加载进 symbol cache - Pro 2.x vs 3.x 是不同存储模型 —— 3.x 走 branch 模型（已打通）， 2.x 走 boards[] 直链（未打通）；判别条件：project meta 的 branch_uuid 是否为 null CLI 新增 --with-pro-source / --backfill-pro-source / --pro-cookie / --origin（按 origin 字段服务端过滤 listing API），crawl_one() 按 origin=pro 自动 dispatch 到 Pro fetcher。 schema：docType 类型从 integer 放宽到 [integer, string, null] （兼容 Std 的 1/3 + Pro 的 BOARD/SCH 等），新增 message_count 字段。 License 注意：本批 5 个项目全是 NC-SA / GPL，未达 Pro source doc §4.2 Forge 白名单（MIT/BSD/Apache/CC0/CC-BY/CERN-OHL-P/Unlicense）。按 CLAUDE.md "研究用、不再分发" 原则 raw 入库无碍；Forge 投影时另过白名单。详细技术细节见 docs/sources/easyeda_pro_source.md rev 3 + log.md。 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>	2026-04-28 21:45:52 +08:00

Author

SHA1

Message

Date

Knowit

eee1a9b97e

crawler: --skip-ext + --max-source-mb gates for batch-50 expansion

Two CLI gates needed before scaling Pro batch beyond top-5:

--skip-ext mp4,qt,mov  (attachment filter)
  Skips video extensions in attachment download. Phase 1 measurements
  showed mp4+qt occupy ~54% of attachment storage. Entry still recorded
  in metadata.json with skipped:ext:<token> so we can re-fetch later if
  the policy changes. Honors both server-declared `ext` and filename
  suffix, case-insensitively.

--max-source-mb N  (Pro source size cap)
  Trips inside the chain replay loop on encrypted-blob total. On trip:
  raise ProjectOversizeError, wipe partial source/, append a row to
  data/state/oshwhub_pro_oversize.jsonl. Lets us shortlist 50+ Pro
  projects without one X86-board-class outlier (~500 MB) blowing the
  LFS budget. Std and Pro 2.x legacy are not capped (both <2 MB in
  sample).

Verified:
  - cap=0 trips on first blob (1.2 MB), source/ wiped, state recorded
  - cap=100 runs full ESP-VoCat (7.5 MB plain, 278 docs)
  - skip-ext microtest: 8/8 cases (case-insensitive, declared/suffix
    fallback, empty-token edge cases)

Plan + frozen candidate list for the next 50 projects:
  - docs/plans/oshwhub_batch50.md
  - data/state/oshwhub_batch50_candidates.jsonl (gitignore exception added)

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-04-29 00:24:55 +08:00

Knowit

1e06ba6582

crawler: split sleep policy by host — chain blob fetches drop 5s -> 0.2s

The Pro modern fetch_pro_modern walks a per-history blob loop on
modules.lceda.cn (CDN-flavored host serving AES-encrypted EPRO2 streams).
We were sleeping 5s between every blob — same rate we use for the
rate-sensitive pro.lceda.cn API host. HAR analysis (proexportNew2.har)
shows the editor fires these blobs back-to-back without throttling, so
0.2s is plenty.

Walltime drops linearly with chain length:
  ESP-VoCat (chain=12):    80s sleep -> 22s sleep  (-72%)
  220V power (chain=28):  160s sleep -> 26s sleep  (-84%)
  X86 board  (chain~700, projection):  ~1h -> ~3min

Verified by re-fetching ESP-VoCat + 220V power: byte-identical output
across all per-doc .epro2 files (sha256 match), only fetched_at
timestamp differs in manifest.json. Two manifest files re-stamped as
proof of the validation runs.

API host sleeps (4x 5s in modern fetcher, 7x 5s in legacy fetcher) are
unchanged — those go to pro.lceda.cn /api/ which still wants polite
QPS<=0.2.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-04-29 00:09:19 +08:00

Knowit

3282a028c4

Add EasyEDA Pro EPRO2 source ingestion (3/5 batch test)

打通 oshwhub origin=pro 现代 Pro 3.x 工程的 EPRO2 源抓取链路。3/5
modern Pro 项目完整解出（共 8423 docs / 542 MB plain）：

- X86 主板        7374 docs / 481 MB plain (chain=85, editor=3.2.15)
- 220V 桌面电源     771 docs /  26 MB plain (chain=28, editor=3.2.69)
- ESP-VoCat       278 docs / 7.5 MB plain (chain=12, editor=3.2.91)

剩余 2/5 是 legacy Pro 2.x（立创泰山派 RK3566、梁山派），项目 meta
返回 branch_uuid=null + editorVersion="2.1.40"，没有 git-style chain
模型，文档直接挂在 boards[].sch/pcb 字段上，访问端点暂未挖通；元
数据落库 metadata.json，source/ 留空。

实现要点：
- fetch_pro_source(): 4 步流程（project → branch HEAD → structures
  → /branches/<B>/histories/<HEAD> 即返完整 chain，无需 ?limit 批量
  端点）+ 逐 history 走 AES-128-GCM 解密（16 字节 IV，pycryptodome
  原生支持）+ gunzip + 按 DOCHEAD 切 per-doc EPRO2 流
- EPRO2 解析坑：行末单 `|` 是行终止符不是字段分隔符，必须先
  rstrip("|") 再 split("||")，否则 payload JSON 解析失败 silently
  swallow 导致 cur_doc 不设 → 第一轮 X86 板 7374 docs 抽出来只剩 2 个
- docType 实测远不止 BOARD/PCB/SCH/SCH_PAGE，还含 SYMBOL /
  FOOTPRINT / DEVICE / BLOB / FONT / CONFIG —— Pro 把组件库快照也
  随项目存到 history，下游做 EPRO2→KiCad 转换时必须先把这些 lib
  doc 加载进 symbol cache
- Pro 2.x vs 3.x 是不同存储模型 —— 3.x 走 branch 模型（已打通），
  2.x 走 boards[] 直链（未打通）；判别条件：project meta 的
  branch_uuid 是否为 null

CLI 新增 --with-pro-source / --backfill-pro-source / --pro-cookie /
--origin（按 origin 字段服务端过滤 listing API），crawl_one() 按
origin=pro 自动 dispatch 到 Pro fetcher。

schema：docType 类型从 integer 放宽到 [integer, string, null]
（兼容 Std 的 1/3 + Pro 的 BOARD/SCH 等），新增 message_count 字段。

License 注意：本批 5 个项目全是 NC-SA / GPL，未达 Pro source doc
§4.2 Forge 白名单（MIT/BSD/Apache/CC0/CC-BY/CERN-OHL-P/Unlicense）。
按 CLAUDE.md "研究用、不再分发" 原则 raw 入库无碍；Forge 投影时
另过白名单。

详细技术细节见 docs/sources/easyeda_pro_source.md rev 3 + log.md。

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-04-28 21:45:52 +08:00

3 Commits