de:wasserversorger-enrich

Journal entry appended. Run complete.

Outcome: Enriched 45 German water suppliers (2 parallel research passes, 23 + 22 companies) — all verified as genuine drinking-water operators with contact fields, emergency numbers, and published water-analysis URLs recorded, none invented. Merged cleanly into the campaign master files: enrichment.jsonl 1929→1974 distinct companies, roster-links.jsonl +242 municipality links, population coverage holds at 99.99% (83.57M of 83.58M). Flagged 8 roster discrepancies for human review (a domain rebrand, two member-list undercounts, a possible domain mismatch on a shared energy/water utility, and a couple of classification questions) plus 4 new host-fetch quirks — all logged in memory/journal.md.

build_enrich_batch_excl.mjs 45 → 0 pending_domain again (roster 3375, enriched 1929 entering, 1466 pending, 6 standing duplicate-mint pairs excluded). Split shard A (23) / shard B (22), dispatched to 2 synchronous foreground Agent calls in parallel, told explicitly not to spawn nested sub-agents and to append after each company. Both ran to completion unattended (161 + 155 tool calls).

Verification before merge: both shard files validated as clean JSON (23 + 22 lines, 0 bad, no duplicate ids), all 45 batch company_ids present across the two outputs, zero collisions with the existing 1929 master records. enrichment.jsonl 1929 → 1974 distinct (+45). map_roster.mjs: 20533 → 20775 links (+242, matching the shards' 116+126 link records exactly). consolidate_roster.mjs: still 177 shards / 3375 distinct companies (unchanged, expected); roster-links.csv 17511 distinct after dedup. coverage.mjs: population 99.99% → 99.99% (unchanged, expected — batch enriches already-claimed companies), same 24 residual municipalities / 11,023 people, Hartmannsdorf-Reichenau still largest at 1,040.

Non-suppliers found: 0 — all 45 confirmed genuine drinking-water operators.

Field recovery (45 companies, both shards combined): name/phone/website_url ~45/45; address 44/45 found (1 not_found); email 42/45 found (2 not_found, 1 not_applicable); short_name 20 found / 25 not_published; contact_form_url 17 found / 5 not_published / 23 not_found; emergency_phone_number 38 found / 3 not_published (earned) / 4 not_found; opening-hours fields weakest again (address: 28 found/17 not_found; phone: 24 found/1 not_applicable/20 not_found); handelsregister_nr mostly not_applicable (Zweckverbände/Eigenbetriebe/KUs), a few genuine finds. Analysis locator: 22/23 (shard A) + 17/22 (shard B) = 39/45 got a real analysis_url; multi_zone true for 14/23 + 12/22 ≈ 58% combined, roughly consistent with the campaign's ~70% prior (a bit lower this batch, no single explanation, not investigated further).

Roster corrections from this batch, flagged for human review:

• wv-de-1762 Stadtwerke Pinneberg: domain fully migrated/rebranded to sw-suedholstein.de (Stadtwerke Südholstein GmbH, HRB 5172) — old stadtwerke-pinneberg.de and its netze. subdomain now redirect/TLS-fail. official_domain left unchanged per the never-silently- overwrite rule; recommend updating it.
• wv-de-828 Zweckverband Wasserversorgung Trollmühle: roster had only 3 municipalities; own site shows a 4-VG, ~43,000-resident, 24-Ortsgemeinde footprint — a major undercount.
• wv-de-10523 Wasserzweckverband Paunzhausen: roster had 8 founding members; own Historie page shows 14 full members plus 5 separate Betriebsführung-only relationships.
• wv-de-16510 Stadtwerke Hemau: roster said Handelsregister "not_checked"; Impressum actually prints HRA 6258 despite being called an Eigenbetrieb.
• wv-de-4305 Markt Elsenfeld: own emergency-numbers page routes water Notdienst to "Zweckverband AMME" (an 18-member regional body), contradicting the roster's "pure self-supply" classification — unresolved, needs a human check, potentially affects other AMME-area companies in the roster.
• wv-de-17014 Gemeindewerke Adelsdorf KU: batch official_domain gw-adelsdorf.de is an energy-only utility (Nahwärme/Windenergie/Strom) with zero water content anywhere in its nav. The actual drinking-water operator is Gemeinde Adelsdorf on the separate domain adelsdorf.de (own Wasserwerk, Uttstadt 16). Domain left unchanged, flagged in both notes and analysis.notes — needs a decision to either repoint the domain or split into two company_ids.
• wv-de-17017 ENTEGA Wasserversorgung Biblis GmbH: confirmed as the correct Biblis-specific subsidiary (HRB 62205), not the parent group — but no standalone Impressum for the GmbH itself was found; address/HR number came from third-party directories (Das Örtliche, Creditsafe), confidence kept medium.
• wv-de-17020 Hohenbrunn: joint Wasserwerk with Gemeinde Ottobrunn — flagged in case Ottobrunn has its own separate company_id elsewhere in the roster (possible double-count risk).
• Several others (Ehringshausen, Schmitten, Schlitz, Wehrheim, Werneuchen, Lügde) had materially bigger/more-structured supply areas (multi-zone Ortsteile, external deliveries) than the roster's self_only/single-member status implied — none contradict self_only at the municipality level, but all understate internal zone structure relevant to the analysis pivot.

New host traps this batch:

• hilzingen.de returns HTTP 403 to the fetch tool on every URL (homepage, water page, Impressum) — recorded as unresolved/medium confidence throughout, not as non-publication.
• aidlingen.de mid-restructure: several previously-recorded deep links (roster's website_url and supply_area_url) now 404; only a legacy ?id=73 URL still resolves.
• reinfeld.de loops infinitely on its index.php?object=... query-string URLs without a cookie jar — curl -c/-b fixed it, plain WebFetch/curl without cookies did not.
• liederbach.eu returned 403 to WebFetch but 200 to a plain curl with a standard browser UA — fetcher-specific, not a real block; don't record as unpublished/inaccessible without trying a UA'd curl first.

For next run: continue with build_enrich_batch_excl.mjs 45. Still 0 pending_domain, 1421 other pending (3375 roster total, 1974 distinct company_ids enriched, 6 excluded). New standing items from this pass: Pinneberg domain rebrand, Trollmühle/Paunzhausen roster undercounts, Hemau HRA number, Elsenfeld/AMME classification question, Adelsdorf domain mismatch, ENTEGA Biblis low-confidence address, Hohenbrunn/Ottobrunn possible double-count. All prior standing items carried in the 2026-09-05 00:30 entry (Malchow's expanded footprint, Walkersbach zone confirmation, two medium-confidence contact-form reads, Heiligenhafen, Perl, Wölfersheim, Lütjenburg HRA, Birkenau/Bad König URLs, Lenggries duplicate-mint, Illergruppe role correction, Kirkel/KEW Neunkirchen, Gommern retry, stadtwerke-mengen.de TLS trap, and the long tail of older items) remain open, none touched this pass.