sync: record fscom targeted verification push

packages/scraper/src/scrapers/fs-com.ts

@@ -72,6 +72,7 @@ const MAX_DETAIL_PAGES_PER_RUN = parseInt(process.env["FS_MAX_DETAIL_PAGES_PER_R
 const STOCK_FRESH_HOURS = parseInt(process.env["FS_STOCK_FRESH_HOURS"] ?? "12", 10);
 const FORCE_REVALIDATE = process.env["TIP_FORCE_REVALIDATE"] === "1";
 const ONLY_MISSING_IMAGES = process.env["FS_ONLY_MISSING_IMAGES"] === "1";
+const DB_DETAIL_ONLY = process.env["FS_DB_DETAIL_ONLY"] === "1";
 
 const PROXY_URLS = (process.env["PROXY_URLS"] ?? "")
   .split(",")
@@ -241,8 +242,18 @@ function detectSpeed(text: string): { speed: string; speedGbps: number } | undef
 }
 
 function detectReach(text: string): string | undefined {
-  const m = text.match(/(\d+)\s*(m|km)\b/i);
-  return m ? `${m[1]}${m[2].toLowerCase()}` : undefined;
+  const m = text.match(/(\d{1,3}(?:,\d{3})+|\d+(?:\.\d+)?)\s*(m|km)\b/i);
+  return m ? `${m[1].replace(/,/g, "")}${m[2].toLowerCase()}` : undefined;
+}
+
+function detectFiberType(text: string): string | undefined {
+  if (/active\s+optical|\baoc\b/i.test(text)) return "AOC";
+  if (/copper|dac|twinax|direct\s+attach|rj-?45|base-t/i.test(text)) return "Copper";
+  if (/single.?mode|\bsmf\b|os2|cwdm|dwdm|\bcw-|^cw-|dw-|bidi|\blx\b|\blr\d*\b|\ber\d*\b|\bzr\d*\b|\bdr\d*\b|\bfr\d*\b|\bpsm\d*\b/i.test(text)) {
+    return "SMF";
+  }
+  if (/multi.?mode|\bmmf\b|om[1-5]|\bsx\b|\bsr\d*\b/i.test(text)) return "MMF";
+  return undefined;
 }
 
 // ── Types ──────────────────────────────────────────────────────────────────────
@@ -777,8 +788,45 @@ export async function scrapeFs(): Promise<void> {
   console.log(`Vendor ID: ${vendorId}`);
 
   // ── Phase 1: Discover product URLs ─────────────────────────────────────────
+  let productMap: Map<string, ProductSummary>;
+  if (DB_DETAIL_ONLY) {
+    console.log("\n[Phase 1] DB detail-only mode — using existing FS.COM product URLs with missing verification signals…");
+    const dbRows = await pool.query(
+      `
+      SELECT t.part_number, t.product_page_url
+      FROM transceivers t
+      JOIN vendors v ON v.id = t.vendor_id
+      WHERE v.name = 'FS.COM'
+        AND t.product_page_url IS NOT NULL
+        AND t.product_page_url != ''
+        AND t.product_page_url LIKE '%/products/%'
+        AND (
+          COALESCE(t.price_verified, false) = false
+          OR COALESCE(t.image_verified, false) = false
+          OR COALESCE(t.details_verified, false) = false
+          OR COALESCE(t.fiber_type, '') = ''
+          OR COALESCE(t.reach_label, '') = ''
+        )
+      ORDER BY
+        COALESCE(t.price_verified, false) DESC,
+        COALESCE(t.image_verified, false) DESC,
+        COALESCE(t.details_verified, false) ASC,
+        t.part_number
+      LIMIT $1
+    `,
+      [MAX_DETAIL_PAGES_PER_RUN]
+    );
+    productMap = new Map(
+      dbRows.rows.map((row) => {
+        const url = normalizeFsProductUrl(row.product_page_url as string);
+        const partNumber = row.part_number as string;
+        return [url, { url, name: partNumber, partNumber }];
+      })
+    );
+  } else {
     console.log("\n[Phase 1] Collecting product URLs from category listing pages…");
-  const productMap = await collectProductUrls(proxyConfiguration);
+    productMap = await collectProductUrls(proxyConfiguration);
+  }
 
   if (productMap.size === 0) {
     console.warn("[Phase 1] No products discovered — check selectors or proxy.");
@@ -860,16 +908,19 @@ export async function scrapeFs(): Promise<void> {
       const speedInfo = detectSpeed(detail.name);
       const reach = detectReach(detail.name);
       const parsed = parseSpecTable(detail.specs);
+      const textForInference = `${detail.name} ${detail.partNumber} ${Object.values(detail.specs).join(" ")}`;
+      const fiberType = parsed.fiberType ?? detectFiberType(textForInference);
 
       const transceiverId = await findOrCreateScrapedTransceiver({
         partNumber: detail.partNumber,
         vendorId,
+        productUrl: detail.url,
         formFactor: ff,
         speedGbps: speedInfo?.speedGbps,
         speed: speedInfo?.speed,
         reachLabel: reach ?? parsed.reachLabel,
         reachMeters: parsed.reachMeters,
-        fiberType: parsed.fiberType,
+        fiberType,
         wavelengths: parsed.wavelengths,
         imageUrl: detail.imageUrl,
         category: "DataCenter",
@@ -922,7 +973,7 @@ export async function scrapeFs(): Promise<void> {
       if (Object.keys(detail.specs).length > 0) {
         const updated = await updateVerifiedSpecs({
           transceiverId,
-          fiberType: parsed.fiberType,
+          fiberType,
           connector: parsed.connector,
           wavelengths: parsed.wavelengths,
           reachMeters: parsed.reachMeters,
@@ -933,7 +984,7 @@ export async function scrapeFs(): Promise<void> {
           domSupport: parsed.domSupport,
           imageUrl: detail.imageUrl,
           datasheetUrl: detail.datasheetUrl,
-          source: "fs.com",
+          source: detail.url,
         });
         if (updated) specsUpdated++;
       }

sync/CURRENT.md

@@ -1,9 +1,96 @@
 # Current TIP Sync State
 
-Updated: 2026-05-09 07:34 UTC
+Updated: 2026-05-09 09:18 UTC
 
 ## Newest Work
 
+- TIP FS.com / Fiberstore targeted verification push on 2026-05-09:
+  - operator requested FS.com/Fiberstore next, with all crawler/scraper/robot learnings written to the TIPLLM training pool and no external AI
+  - code improved:
+    - `packages/scraper/src/scrapers/fs-com.ts`
+      - added `FS_DB_DETAIL_ONLY=1` mode to revalidate existing FS.COM product URLs directly from DB
+      - avoids broad category/listing discovery while product URLs still need verification
+      - `detectReach()` now handles comma thousands and decimal values
+      - added deterministic `detectFiberType()` fallback from product name, part number and specs
+      - scraper now writes `productUrl` into the transceiver row
+      - detail verification source is now the actual FS.com product URL instead of the literal `fs.com`
+  - live Erik verification:
+    - deployed scraper to `/opt/tip`
+    - `pnpm -C packages/scraper build` passed on Erik after the change
+    - ran four safe DB-detail-only Playwright batches:
+      - batch 1: target `80`, scraped `80`, failed `0`, new prices `17`, stock `18`, specs `24`
+      - batch 2: target `80`, scraped `79`, failed `0`, new prices `6`, stock `8`, specs `23`
+      - batch 3: target `90`, scraped `89`, failed `0`, new prices `21`, stock `24`, specs `47`
+      - batch 4 closure: target `42`, scraped `42`, failed `0`, new prices `5`, stock `3`, specs `25`
+    - all runs used Playwright concurrency `1`, `nice -n 10`, and no broad category crawl
+    - Erik/TIP health after closure:
+      - status: `healthy`
+      - load status: `ok`
+      - memory used: `13%`
+      - transceivers: `17647`
+      - vendors: `478`
+      - switches: `680`
+      - global verified counters:
+        - price: `11557`
+        - image: `10636`
+        - details: `9816`
+        - fully: `8522`
+  - FS.com before targeted detail batches:
+    - total rows: `383`
+    - price verified: `379`
+    - image verified: `299`
+    - details verified: `108`
+    - price+image+details: `108`
+    - fully verified: `3`
+    - missing product URL: `76`
+    - missing image URL: `84`
+    - missing reach label: `9`
+    - missing fiber type: `323`
+    - HTML product-like complete rows: `106`
+  - FS.com after closure:
+    - total rows: `383`
+    - price verified: `379`
+    - image verified: `299`
+    - details verified: `260`
+    - price+image+details: `260`
+    - fully verified: `205`
+    - missing product URL: `76`
+    - missing image URL: `84`
+    - missing reach label: `9`
+    - missing fiber type: `123`
+    - HTML product-like rows:
+      - total `299`
+      - price `299`
+      - image `282`
+      - details `258`
+      - complete `258`
+    - no-url rows:
+      - total `76`
+      - price `76`
+      - image `15`
+      - details `0`
+    - category rows:
+      - total `4`
+      - no verified signals
+  - interpretation / next strategy:
+    - the DB-detail-only approach is now mostly exhausted
+    - the fourth clean closure batch did not raise `details_verified`; it only nudged `fully_verified` from `199` to `205`
+    - do not keep repeating the same FS.com detail crawler on Erik
+    - next FS.com work should be:
+      - source-discovery/classification robot for the `76` no-url rows
+      - parser/source diagnostics for the remaining `41` HTML product-like rows missing detail/fiber/image signals
+      - likely separate handling for malformed or historical `/de/de/products/...` URLs and pages that return no useful text
+  - TIPLLM training pool:
+    - all four FS.com batches were written and pushed to Gitea
+    - latest training commits:
+      - `28cac05` batch 1
+      - `a0a6be3` batch 2
+      - `38736ae` batch 3
+      - `2c25bf3` closure batch
+  - important truth:
+    - do not claim FS.com is complete
+    - the honest current claim is: FS.com product-like coverage improved strongly, but `258/299` HTML product-like rows are complete and `76` no-url rows still need source discovery/classification
+
 - TIP Flexoptix completion push on 2026-05-09:
   - operator said "feuer frei" after confirming Flexoptix was not yet complete
   - TIPLLM training pool was updated immediately with the truth rule:

sync/history/2026-05-09-fscom-targeted-verification-push.md

@@ -0,0 +1,101 @@
+# FS.com / Fiberstore Targeted Verification Push
+
+Date: 2026-05-09
+
+## Intent
+
+Continue TIP data completion for FS.com/Fiberstore after Flexoptix. The operator requested price, image and product information to be researched deeply enough to avoid manual validation, while keeping Erik safe and writing every crawler/scraper/robot learning into the TIPLLM training pool.
+
+## Code Changed
+
+- `packages/scraper/src/scrapers/fs-com.ts`
+  - added `FS_DB_DETAIL_ONLY=1`
+  - targets existing FS.COM DB product URLs with missing verification signals
+  - avoids broad category discovery while known product URLs still need work
+  - improved reach parsing for comma/decimal values
+  - added deterministic fiber type fallback from product name, part number and specs
+  - writes product URL to `transceivers.product_page_url`
+  - stores the real FS.com product URL as detail verification source
+
+## Live Runs
+
+All runs were on Erik with:
+
+- Playwright concurrency `1`
+- `nice -n 10`
+- no broad category crawl
+- DB-detail-only mode
+
+Batch results:
+
+- Batch 1: target `80`, scraped `80`, failed `0`, new prices `17`, stock `18`, specs `24`
+- Batch 2: target `80`, scraped `79`, failed `0`, new prices `6`, stock `8`, specs `23`
+- Batch 3: target `90`, scraped `89`, failed `0`, new prices `21`, stock `24`, specs `47`
+- Batch 4 closure: target `42`, scraped `42`, failed `0`, new prices `5`, stock `3`, specs `25`
+
+`pnpm -C packages/scraper build` passed on Erik after the scraper change.
+
+## FS.com Counters
+
+Before:
+
+- total rows: `383`
+- price verified: `379`
+- image verified: `299`
+- details verified: `108`
+- price+image+details: `108`
+- fully verified: `3`
+- missing URL: `76`
+- missing image URL: `84`
+- missing reach label: `9`
+- missing fiber type: `323`
+- HTML product-like complete: `106`
+
+After closure:
+
+- total rows: `383`
+- price verified: `379`
+- image verified: `299`
+- details verified: `260`
+- price+image+details: `260`
+- fully verified: `205`
+- missing URL: `76`
+- missing image URL: `84`
+- missing reach label: `9`
+- missing fiber type: `123`
+- HTML product-like rows: `299`
+- HTML product-like complete: `258`
+- no-url rows: `76`
+- category rows: `4`
+
+TIP health after closure:
+
+- status: `healthy`
+- load status: `ok`
+- memory used: `13%`
+- transceivers: `17647`
+- vendors: `478`
+- switches: `680`
+- fully verified globally: `8522`
+
+## Training Pool
+
+FS.com batches were written to `/tmp/tip-training-data` and pushed to Gitea.
+
+Training pool commits:
+
+- `28cac05 crawl: add fscom db detail batch learning record`
+- `a0a6be3 crawl: add fscom db detail batch 2 learning record`
+- `38736ae crawl: add fscom db detail batch 3 learning record`
+- `2c25bf3 crawl: add fscom db detail closure learning record`
+
+## Next
+
+Do not repeat the same DB-detail-only FS.com crawler on Erik. The fourth clean closure batch did not increase `details_verified`, so the remaining gaps need a different strategy:
+
+- source-discovery/classification for `76` no-url rows
+- parser/source diagnostics for the remaining `41` HTML product-like rows missing details or fiber/image signals
+- explicit classification for `4` category rows
+- likely cleanup of historical/malformed `/de/de/products/...` URLs and no-text pages
+
+Truth rule: do not claim FS.com is complete. Current honest status is `258/299` HTML product-like rows complete and `205/383` fully verified overall.