riksdagsmonitor 0.8.12 → 0.8.15

package/dist/scripts/populate-analysis-data.js

@@ -0,0 +1,255 @@
+#!/usr/bin/env tsx
+/**
+ * @module populate-analysis-data
+ * @description Standalone script that fetches recent data from **all** MCP sources
+ * (riksdag-regering, SCB, World Bank) and persists it to `analysis/data/`.
+ *
+ * This populates the analysis folder with **all** supported data types:
+ * - **Documents**: propositions, motions, committeeReports, votes, speeches,
+ *   questions, interpellations
+ * - **Calendar events**: upcoming parliamentary events
+ * - **MPs**: current member profiles
+ * - **Government documents**: recent government publications (regeringen.se)
+ * - **Voting groups**: party-level voting patterns
+ * - **World Bank indicators**: Swedish economic context (GDP, unemployment, etc.)
+ * - **SCB statistics**: Swedish official statistics for key policy domains
+ *
+ * The script reuses the existing MCP client and data-persistence modules,
+ * ensuring the collision-free sidecar design (data + .meta.json) is
+ * consistently applied.
+ *
+ * Usage:
+ *   npx tsx scripts/populate-analysis-data.ts [--limit N] [--date YYYY-MM-DD]
+ *
+ * Options:
+ *   --limit N           Max documents per type (default: 20)
+ *   --date YYYY-MM-DD   Target date for analysis (default: today UTC)
+ *
+ * @author Hack23 AB
+ * @license Apache-2.0
+ */
+import { MCPClient } from './mcp-client/client.js';
+import { downloadAllDocuments, flattenDocuments, } from './pre-article-analysis/data-downloader.js';
+import { persistDownloadedData, persistEvents, persistMPs, persistMCPResponse, persistWorldBankData, persistSCBData, getDataRoot, resolveDocId, } from './pre-article-analysis/data-persistence.js';
+import { WorldBankClient, INDICATOR_IDS, COUNTRY_CODES, } from './world-bank-client.js';
+import { SCBClient, SCB_DOMAINS, } from './scb-client.js';
+// ---------------------------------------------------------------------------
+// Helpers
+// ---------------------------------------------------------------------------
+/**
+ * Compute riksmöte (parliamentary session year) from an ISO date string.
+ * Swedish parliamentary year runs Oct → Sep:
+ *   - 2025-10-01 to 2026-09-30 → "2025/26"
+ *   - 2026-10-01 to 2027-09-30 → "2026/27"
+ *
+ * @param dateStr - ISO date string (YYYY-MM-DD)
+ * @returns Riksmöte identifier (e.g. "2025/26")
+ */
+export function riksMoteFromDate(dateStr) {
+    const d = new Date(dateStr + 'T00:00:00Z');
+    const year = d.getUTCFullYear();
+    const month = d.getUTCMonth() + 1;
+    if (month >= 10)
+        return `${year}/${String(year + 1).slice(-2)}`;
+    return `${year - 1}/${String(year).slice(-2)}`;
+}
+/** Format a Date to YYYY-MM-DD. */
+function formatDate(d) {
+    return d.toISOString().slice(0, 10);
+}
+/** Safely extract error message from unknown thrown value. */
+function errorMessage(err) {
+    return err instanceof Error ? err.message : String(err);
+}
+function parseArgs() {
+    const args = process.argv.slice(2);
+    let limit = 20;
+    let date = formatDate(new Date());
+    for (let i = 0; i < args.length; i++) {
+        if (args[i] === '--limit' && args[i + 1]) {
+            limit = Math.max(1, parseInt(args[i + 1], 10) || 20);
+            i++;
+        }
+        else if (args[i] === '--date' && args[i + 1]) {
+            const candidate = args[i + 1];
+            const dm = /^(\d{4})-(\d{2})-(\d{2})$/.exec(candidate);
+            if (!dm) {
+                console.error(`❌ Invalid date: "${candidate}". Expected YYYY-MM-DD format.`);
+                process.exit(1);
+            }
+            const [, yStr, mStr, dStr] = dm;
+            const y = Number(yStr), mo = Number(mStr), da = Number(dStr);
+            const parsed = new Date(Date.UTC(y, mo - 1, da));
+            if (Number.isNaN(parsed.getTime()) || parsed.getUTCFullYear() !== y || parsed.getUTCMonth() + 1 !== mo || parsed.getUTCDate() !== da) {
+                console.error(`❌ Invalid date: "${candidate}". Expected a valid YYYY-MM-DD date.`);
+                process.exit(1);
+            }
+            date = candidate;
+            i++;
+        }
+    }
+    return { limit, date };
+}
+// ---------------------------------------------------------------------------
+// Main
+// ---------------------------------------------------------------------------
+async function main() {
+    const { limit, date } = parseArgs();
+    const rm = riksMoteFromDate(date);
+    console.log('╔══════════════════════════════════════════════════════════════╗');
+    console.log('║   📥 Populate Analysis Data — MCP Data Fetcher             ║');
+    console.log('╚══════════════════════════════════════════════════════════════╝');
+    console.log(`   📅 Date: ${date}`);
+    console.log(`   🏛️  Riksmöte: ${rm}`);
+    console.log(`   📊 Limit: ${limit} per type`);
+    console.log(`   📂 Data root: ${getDataRoot()}`);
+    console.log('');
+    const client = new MCPClient();
+    // ── Step 1: Download all document types ─────────────────────────────────
+    console.log('📄 Step 1/7: Downloading parliamentary documents...');
+    try {
+        const { data, manifest } = await downloadAllDocuments(client, { limit, rm });
+        const allDocs = flattenDocuments(data);
+        console.log(`   ✅ Downloaded ${allDocs.length} documents from ${manifest.dataSources.length} MCP tools (${manifest.durationMs}ms)`);
+        const persistResult = persistDownloadedData(data, rm);
+        console.log(`   🗄️  Persisted ${persistResult.written} document files (${persistResult.skipped} skipped)`);
+    }
+    catch (err) {
+        console.error('   ❌ Document download failed:', errorMessage(err));
+    }
+    // ── Step 2: Download calendar events ────────────────────────────────────
+    console.log('\n📅 Step 2/7: Downloading calendar events...');
+    try {
+        const today = new Date(date + 'T00:00:00Z');
+        const twoWeeksAhead = new Date(today);
+        twoWeeksAhead.setUTCDate(twoWeeksAhead.getUTCDate() + 14);
+        const rawEvents = await client.fetchCalendarEvents(formatDate(today), formatDate(twoWeeksAhead));
+        const events = (Array.isArray(rawEvents) ? rawEvents : []);
+        console.log(`   ✅ Downloaded ${events.length} calendar events`);
+        if (events.length > 0) {
+            const persistResult = persistEvents(events, rm);
+            console.log(`   🗄️  Persisted ${persistResult.written} event files (${persistResult.skipped} skipped)`);
+        }
+    }
+    catch (err) {
+        console.error('   ❌ Calendar event download failed:', errorMessage(err));
+    }
+    // ── Step 3: Download MP profiles ────────────────────────────────────────
+    console.log('\n👤 Step 3/7: Downloading MP profiles...');
+    try {
+        const rawMPs = await client.fetchMPs({ limit });
+        const mps = (Array.isArray(rawMPs) ? rawMPs : []);
+        console.log(`   ✅ Downloaded ${mps.length} MP profiles`);
+        if (mps.length > 0) {
+            const persistResult = persistMPs(mps, rm);
+            console.log(`   🗄️  Persisted ${persistResult.written} MP files (${persistResult.skipped} skipped)`);
+        }
+    }
+    catch (err) {
+        console.error('   ❌ MP download failed:', errorMessage(err));
+    }
+    // ── Step 4: Download government documents ───────────────────────────────
+    console.log('\n🏛️  Step 4/7: Downloading government documents...');
+    try {
+        const thirtyDaysAgo = new Date(date + 'T00:00:00Z');
+        thirtyDaysAgo.setUTCDate(thirtyDaysAgo.getUTCDate() - 30);
+        const rawGovDocs = await client.fetchGovernmentDocuments({
+            dateFrom: formatDate(thirtyDaysAgo),
+            dateTo: date,
+            limit,
+        });
+        const govDocs = (Array.isArray(rawGovDocs) ? rawGovDocs : []);
+        console.log(`   ✅ Downloaded ${govDocs.length} government documents`);
+        let govWritten = 0;
+        for (let i = 0; i < govDocs.length; i++) {
+            const doc = govDocs[i];
+            if (!doc)
+                continue;
+            persistMCPResponse({ tool: 'search_regering', params: { dateFrom: formatDate(thirtyDaysAgo), dateTo: date, limit }, server: 'riksdag-regering' }, doc, resolveDocId(doc, i));
+            govWritten++;
+        }
+        console.log(`   🗄️  Persisted ${govWritten} government document files`);
+    }
+    catch (err) {
+        console.error('   ❌ Government document download failed:', errorMessage(err));
+    }
+    // ── Step 5: Download voting groups by party ─────────────────────────────
+    console.log('\n🗳️  Step 5/7: Downloading voting groups by party...');
+    try {
+        const rawGroups = await client.fetchVotingGroup({ rm, groupBy: 'parti', limit });
+        const groups = (Array.isArray(rawGroups) ? rawGroups : []);
+        console.log(`   ✅ Downloaded ${groups.length} voting group records`);
+        let voteGroupWritten = 0;
+        for (let i = 0; i < groups.length; i++) {
+            const group = groups[i];
+            if (!group)
+                continue;
+            const record = group;
+            const id = (typeof record['parti'] === 'string' && record['parti'])
+                || `group-${i + 1}`;
+            persistMCPResponse({ tool: 'get_voting_group', params: { rm, groupBy: 'parti', limit }, server: 'riksdag-regering' }, group, `${rm.replace('/', '-')}-${id}`);
+            voteGroupWritten++;
+        }
+        console.log(`   🗄️  Persisted ${voteGroupWritten} voting group files`);
+    }
+    catch (err) {
+        console.error('   ❌ Voting group download failed:', errorMessage(err));
+    }
+    // ── Step 6: Download World Bank indicators for Sweden ───────────────────
+    console.log('\n🌍 Step 6/7: Downloading World Bank indicators for Sweden...');
+    try {
+        const wb = new WorldBankClient();
+        const indicatorEntries = Object.entries(INDICATOR_IDS);
+        let wbWritten = 0;
+        let wbFailed = 0;
+        for (const [_name, indicatorId] of indicatorEntries) {
+            try {
+                const dataPoints = await wb.getIndicator(COUNTRY_CODES.sweden, indicatorId, 10);
+                if (dataPoints.length > 0) {
+                    persistWorldBankData(indicatorId, COUNTRY_CODES.sweden, dataPoints);
+                    wbWritten++;
+                }
+            }
+            catch {
+                wbFailed++;
+            }
+        }
+        console.log(`   ✅ Persisted ${wbWritten}/${indicatorEntries.length} World Bank indicators (${wbFailed} failed)`);
+    }
+    catch (err) {
+        console.error('   ❌ World Bank data download failed:', errorMessage(err));
+    }
+    // ── Step 7: Download SCB statistics for key domains ─────────────────────
+    console.log('\n📊 Step 7/7: Downloading SCB statistics...');
+    try {
+        const scb = new SCBClient();
+        const domainsWithTables = SCB_DOMAINS.filter(d => d.tables.length > 0);
+        let scbWritten = 0;
+        let scbFailed = 0;
+        for (const domain of domainsWithTables) {
+            for (const tableId of domain.tables) {
+                try {
+                    const data = await scb.getTableData(tableId);
+                    if (data && (Array.isArray(data) ? data.length > 0 : true)) {
+                        persistSCBData(tableId, data, { domain: domain.domain, query: domain.query });
+                        scbWritten++;
+                    }
+                }
+                catch {
+                    scbFailed++;
+                }
+            }
+        }
+        console.log(`   ✅ Persisted ${scbWritten} SCB tables (${scbFailed} failed)`);
+    }
+    catch (err) {
+        console.error('   ❌ SCB data download failed:', errorMessage(err));
+    }
+    console.log('\n✅ Analysis data population complete.');
+    console.log(`   📂 Data written to: ${getDataRoot()}`);
+}
+main().catch((err) => {
+    console.error('Fatal error:', err);
+    process.exit(1);
+});
+//# sourceMappingURL=populate-analysis-data.js.map

package/dist/scripts/populate-analysis-data.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"populate-analysis-data.js","sourceRoot":"","sources":["../../scripts/populate-analysis-data.ts"],"names":[],"mappings":";AACA;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AAEH,OAAO,EAAE,SAAS,EAAE,MAAM,wBAAwB,CAAC;AACnD,OAAO,EACL,oBAAoB,EACpB,gBAAgB,GACjB,MAAM,2CAA2C,CAAC;AACnD,OAAO,EACL,qBAAqB,EACrB,aAAa,EACb,UAAU,EACV,kBAAkB,EAClB,oBAAoB,EACpB,cAAc,EACd,WAAW,EACX,YAAY,GACb,MAAM,4CAA4C,CAAC;AAEpD,OAAO,EACL,eAAe,EACf,aAAa,EACb,aAAa,GACd,MAAM,wBAAwB,CAAC;AAChC,OAAO,EACL,SAAS,EACT,WAAW,GACZ,MAAM,iBAAiB,CAAC;AAEzB,8EAA8E;AAC9E,UAAU;AACV,8EAA8E;AAE9E;;;;;;;;GAQG;AACH,MAAM,UAAU,gBAAgB,CAAC,OAAe;IAC9C,MAAM,CAAC,GAAG,IAAI,IAAI,CAAC,OAAO,GAAG,YAAY,CAAC,CAAC;IAC3C,MAAM,IAAI,GAAG,CAAC,CAAC,cAAc,EAAE,CAAC;IAChC,MAAM,KAAK,GAAG,CAAC,CAAC,WAAW,EAAE,GAAG,CAAC,CAAC;IAClC,IAAI,KAAK,IAAI,EAAE;QAAE,OAAO,GAAG,IAAI,IAAI,MAAM,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IAChE,OAAO,GAAG,IAAI,GAAG,CAAC,IAAI,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;AACjD,CAAC;AAED,mCAAmC;AACnC,SAAS,UAAU,CAAC,CAAO;IACzB,OAAO,CAAC,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;AACtC,CAAC;AAED,8DAA8D;AAC9D,SAAS,YAAY,CAAC,GAAY;IAChC,OAAO,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;AAC1D,CAAC;AAED,SAAS,SAAS;IAChB,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IACnC,IAAI,KAAK,GAAG,EAAE,CAAC;IACf,IAAI,IAAI,GAAG,UAAU,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC;IAElC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACrC,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,SAAS,IAAI,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;YACzC,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,QAAQ,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC;YACrD,CAAC,EAAE,CAAC;QACN,CAAC;aAAM,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,QAAQ,IAAI,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;YAC/C,MAAM,SAAS,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YAC9B,MAAM,EAAE,GAAG,2BAA2B,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;YACvD,IAAI,CAAC,EAAE,EAAE,CAAC;gBACR,OAAO,CAAC,KAAK,CAAC,oBAAoB,SAAS,gCAAgC,CAAC,CAAC;gBAC7E,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YAClB,CAAC;YACD,MAAM,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,GAAG,EAAE,CAAC;YAChC,MAAM,CAAC,GAAG,MAAM,CAAC,IAAI,CAAC,EAAE,EAAE,GAAG,MAAM,CAAC,IAAI,CAAC,EAAE,EAAE,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC;YAC7D,MAAM,MAAM,GAAG,IAAI,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,GAAG,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC;YACjD,IAAI,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC,IAAI,MAAM,CAAC,cAAc,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,WAAW,EAAE,GAAG,CAAC,KAAK,EAAE,IAAI,MAAM,CAAC,UAAU,EAAE,KAAK,EAAE,EAAE,CAAC;gBACrI,OAAO,CAAC,KAAK,CAAC,oBAAoB,SAAS,sCAAsC,CAAC,CAAC;gBACnF,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YAClB,CAAC;YACD,IAAI,GAAG,SAAS,CAAC;YACjB,CAAC,EAAE,CAAC;QACN,CAAC;IACH,CAAC;IACD,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC;AACzB,CAAC;AAED,8EAA8E;AAC9E,OAAO;AACP,8EAA8E;AAE9E,KAAK,UAAU,IAAI;IACjB,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,GAAG,SAAS,EAAE,CAAC;IACpC,MAAM,EAAE,GAAG,gBAAgB,CAAC,IAAI,CAAC,CAAC;IAElC,OAAO,CAAC,GAAG,CAAC,kEAAkE,CAAC,CAAC;IAChF,OAAO,CAAC,GAAG,CAAC,gEAAgE,CAAC,CAAC;IAC9E,OAAO,CAAC,GAAG,CAAC,kEAAkE,CAAC,CAAC;IAChF,OAAO,CAAC,GAAG,CAAC,eAAe,IAAI,EAAE,CAAC,CAAC;IACnC,OAAO,CAAC,GAAG,CAAC,qBAAqB,EAAE,EAAE,CAAC,CAAC;IACvC,OAAO,CAAC,GAAG,CAAC,gBAAgB,KAAK,WAAW,CAAC,CAAC;IAC9C,OAAO,CAAC,GAAG,CAAC,oBAAoB,WAAW,EAAE,EAAE,CAAC,CAAC;IACjD,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;IAEhB,MAAM,MAAM,GAAG,IAAI,SAAS,EAAE,CAAC;IAE/B,2EAA2E;IAC3E,OAAO,CAAC,GAAG,CAAC,qDAAqD,CAAC,CAAC;IACnE,IAAI,CAAC;QACH,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,GAAG,MAAM,oBAAoB,CAAC,MAAM,EAAE,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,CAAC;QAC7E,MAAM,OAAO,GAAG,gBAAgB,CAAC,IAAI,CAAC,CAAC;QACvC,OAAO,CAAC,GAAG,CAAC,mBAAmB,OAAO,CAAC,MAAM,mBAAmB,QAAQ,CAAC,WAAW,CAAC,MAAM,eAAe,QAAQ,CAAC,UAAU,KAAK,CAAC,CAAC;QAEpI,MAAM,aAAa,GAAG,qBAAqB,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;QACtD,OAAO,CAAC,GAAG,CAAC,qBAAqB,aAAa,CAAC,OAAO,oBAAoB,aAAa,CAAC,OAAO,WAAW,CAAC,CAAC;IAC9G,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,OAAO,CAAC,KAAK,CAAC,gCAAgC,EAAE,YAAY,CAAC,GAAG,CAAC,CAAC,CAAC;IACrE,CAAC;IAED,2EAA2E;IAC3E,OAAO,CAAC,GAAG,CAAC,+CAA+C,CAAC,CAAC;IAC7D,IAAI,CAAC;QACH,MAAM,KAAK,GAAG,IAAI,IAAI,CAAC,IAAI,GAAG,YAAY,CAAC,CAAC;QAC5C,MAAM,aAAa,GAAG,IAAI,IAAI,CAAC,KAAK,CAAC,CAAC;QACtC,aAAa,CAAC,UAAU,CAAC,aAAa,CAAC,UAAU,EAAE,GAAG,EAAE,CAAC,CAAC;QAE1D,MAAM,SAAS,GAAG,MAAM,MAAM,CAAC,mBAAmB,CAChD,UAAU,CAAC,KAAK,CAAC,EACjB,UAAU,CAAC,aAAa,CAAC,CAC1B,CAAC;QACF,MAAM,MAAM,GAAkB,CAAC,KAAK,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,CAAkB,CAAC;QAC3F,OAAO,CAAC,GAAG,CAAC,mBAAmB,MAAM,CAAC,MAAM,kBAAkB,CAAC,CAAC;QAEhE,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACtB,MAAM,aAAa,GAAG,aAAa,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;YAChD,OAAO,CAAC,GAAG,CAAC,qBAAqB,aAAa,CAAC,OAAO,iBAAiB,aAAa,CAAC,OAAO,WAAW,CAAC,CAAC;QAC3G,CAAC;IACH,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,OAAO,CAAC,KAAK,CAAC,sCAAsC,EAAE,YAAY,CAAC,GAAG,CAAC,CAAC,CAAC;IAC3E,CAAC;IAED,2EAA2E;IAC3E,OAAO,CAAC,GAAG,CAAC,2CAA2C,CAAC,CAAC;IACzD,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,MAAM,MAAM,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC;QAChD,MAAM,GAAG,GAAkB,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAkB,CAAC;QAClF,OAAO,CAAC,GAAG,CAAC,mBAAmB,GAAG,CAAC,MAAM,cAAc,CAAC,CAAC;QAEzD,IAAI,GAAG,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACnB,MAAM,aAAa,GAAG,UAAU,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;YAC1C,OAAO,CAAC,GAAG,CAAC,qBAAqB,aAAa,CAAC,OAAO,cAAc,aAAa,CAAC,OAAO,WAAW,CAAC,CAAC;QACxG,CAAC;IACH,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,OAAO,CAAC,KAAK,CAAC,0BAA0B,EAAE,YAAY,CAAC,GAAG,CAAC,CAAC,CAAC;IAC/D,CAAC;IAED,2EAA2E;IAC3E,OAAO,CAAC,GAAG,CAAC,sDAAsD,CAAC,CAAC;IACpE,IAAI,CAAC;QACH,MAAM,aAAa,GAAG,IAAI,IAAI,CAAC,IAAI,GAAG,YAAY,CAAC,CAAC;QACpD,aAAa,CAAC,UAAU,CAAC,aAAa,CAAC,UAAU,EAAE,GAAG,EAAE,CAAC,CAAC;QAE1D,MAAM,UAAU,GAAG,MAAM,MAAM,CAAC,wBAAwB,CAAC;YACvD,QAAQ,EAAE,UAAU,CAAC,aAAa,CAAC;YACnC,MAAM,EAAE,IAAI;YACZ,KAAK;SACN,CAAC,CAAC;QACH,MAAM,OAAO,GAAkB,CAAC,KAAK,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,EAAE,CAAkB,CAAC;QAC9F,OAAO,CAAC,GAAG,CAAC,mBAAmB,OAAO,CAAC,MAAM,uBAAuB,CAAC,CAAC;QAEtE,IAAI,UAAU,GAAG,CAAC,CAAC;QACnB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACxC,MAAM,GAAG,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;YACvB,IAAI,CAAC,GAAG;gBAAE,SAAS;YACnB,kBAAkB,CAChB,EAAE,IAAI,EAAE,iBAAiB,EAAE,MAAM,EAAE,EAAE,QAAQ,EAAE,UAAU,CAAC,aAAa,CAAC,EAAE,MAAM,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,MAAM,EAAE,kBAAkB,EAAE,EAC7H,GAAG,EACH,YAAY,CAAC,GAAG,EAAE,CAAC,CAAC,CACrB,CAAC;YACF,UAAU,EAAE,CAAC;QACf,CAAC;QACD,OAAO,CAAC,GAAG,CAAC,qBAAqB,UAAU,4BAA4B,CAAC,CAAC;IAC3E,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,OAAO,CAAC,KAAK,CAAC,2CAA2C,EAAE,YAAY,CAAC,GAAG,CAAC,CAAC,CAAC;IAChF,CAAC;IAED,2EAA2E;IAC3E,OAAO,CAAC,GAAG,CAAC,wDAAwD,CAAC,CAAC;IACtE,IAAI,CAAC;QACH,MAAM,SAAS,GAAG,MAAM,MAAM,CAAC,gBAAgB,CAAC,EAAE,EAAE,EAAE,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC,CAAC;QACjF,MAAM,MAAM,GAAkB,CAAC,KAAK,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,CAAkB,CAAC;QAC3F,OAAO,CAAC,GAAG,CAAC,mBAAmB,MAAM,CAAC,MAAM,uBAAuB,CAAC,CAAC;QAErE,IAAI,gBAAgB,GAAG,CAAC,CAAC;QACzB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACvC,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;YACxB,IAAI,CAAC,KAAK;gBAAE,SAAS;YACrB,MAAM,MAAM,GAAG,KAAgC,CAAC;YAChD,MAAM,EAAE,GAAG,CAAC,OAAO,MAAM,CAAC,OAAO,CAAC,KAAK,QAAQ,IAAI,MAAM,CAAC,OAAO,CAAC,CAAC;mBAC9D,SAAS,CAAC,GAAG,CAAC,EAAE,CAAC;YACtB,kBAAkB,CAChB,EAAE,IAAI,EAAE,kBAAkB,EAAE,MAAM,EAAE,EAAE,EAAE,EAAE,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,EAAE,MAAM,EAAE,kBAAkB,EAAE,EACjG,KAAK,EACL,GAAG,EAAE,CAAC,OAAO,CAAC,GAAG,EAAE,GAAG,CAAC,IAAI,EAAE,EAAE,CAChC,CAAC;YACF,gBAAgB,EAAE,CAAC;QACrB,CAAC;QACD,OAAO,CAAC,GAAG,CAAC,qBAAqB,gBAAgB,qBAAqB,CAAC,CAAC;IAC1E,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,OAAO,CAAC,KAAK,CAAC,oCAAoC,EAAE,YAAY,CAAC,GAAG,CAAC,CAAC,CAAC;IACzE,CAAC;IAED,2EAA2E;IAC3E,OAAO,CAAC,GAAG,CAAC,gEAAgE,CAAC,CAAC;IAC9E,IAAI,CAAC;QACH,MAAM,EAAE,GAAG,IAAI,eAAe,EAAE,CAAC;QACjC,MAAM,gBAAgB,GAAG,MAAM,CAAC,OAAO,CAAC,aAAa,CAAC,CAAC;QACvD,IAAI,SAAS,GAAG,CAAC,CAAC;QAClB,IAAI,QAAQ,GAAG,CAAC,CAAC;QAEjB,KAAK,MAAM,CAAC,KAAK,EAAE,WAAW,CAAC,IAAI,gBAAgB,EAAE,CAAC;YACpD,IAAI,CAAC;gBACH,MAAM,UAAU,GAAG,MAAM,EAAE,CAAC,YAAY,CAAC,aAAa,CAAC,MAAM,EAAE,WAAW,EAAE,EAAE,CAAC,CAAC;gBAChF,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBAC1B,oBAAoB,CAAC,WAAW,EAAE,aAAa,CAAC,MAAM,EAAE,UAAU,CAAC,CAAC;oBACpE,SAAS,EAAE,CAAC;gBACd,CAAC;YACH,CAAC;YAAC,MAAM,CAAC;gBACP,QAAQ,EAAE,CAAC;YACb,CAAC;QACH,CAAC;QACD,OAAO,CAAC,GAAG,CAAC,kBAAkB,SAAS,IAAI,gBAAgB,CAAC,MAAM,2BAA2B,QAAQ,UAAU,CAAC,CAAC;IACnH,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,OAAO,CAAC,KAAK,CAAC,uCAAuC,EAAE,YAAY,CAAC,GAAG,CAAC,CAAC,CAAC;IAC5E,CAAC;IAED,2EAA2E;IAC3E,OAAO,CAAC,GAAG,CAAC,8CAA8C,CAAC,CAAC;IAC5D,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,IAAI,SAAS,EAAE,CAAC;QAC5B,MAAM,iBAAiB,GAAG,WAAW,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QACvE,IAAI,UAAU,GAAG,CAAC,CAAC;QACnB,IAAI,SAAS,GAAG,CAAC,CAAC;QAElB,KAAK,MAAM,MAAM,IAAI,iBAAiB,EAAE,CAAC;YACvC,KAAK,MAAM,OAAO,IAAI,MAAM,CAAC,MAAM,EAAE,CAAC;gBACpC,IAAI,CAAC;oBACH,MAAM,IAAI,GAAG,MAAM,GAAG,CAAC,YAAY,CAAC,OAAO,CAAC,CAAC;oBAC7C,IAAI,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC;wBAC3D,cAAc,CAAC,OAAO,EAAE,IAAI,EAAE,EAAE,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,KAAK,EAAE,MAAM,CAAC,KAAK,EAAE,CAAC,CAAC;wBAC9E,UAAU,EAAE,CAAC;oBACf,CAAC;gBACH,CAAC;gBAAC,MAAM,CAAC;oBACP,SAAS,EAAE,CAAC;gBACd,CAAC;YACH,CAAC;QACH,CAAC;QACD,OAAO,CAAC,GAAG,CAAC,kBAAkB,UAAU,gBAAgB,SAAS,UAAU,CAAC,CAAC;IAC/E,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,OAAO,CAAC,KAAK,CAAC,gCAAgC,EAAE,YAAY,CAAC,GAAG,CAAC,CAAC,CAAC;IACrE,CAAC;IAED,OAAO,CAAC,GAAG,CAAC,wCAAwC,CAAC,CAAC;IACtD,OAAO,CAAC,GAAG,CAAC,0BAA0B,WAAW,EAAE,EAAE,CAAC,CAAC;AACzD,CAAC;AAED,IAAI,EAAE,CAAC,KAAK,CAAC,CAAC,GAAG,EAAE,EAAE;IACnB,OAAO,CAAC,KAAK,CAAC,cAAc,EAAE,GAAG,CAAC,CAAC;IACnC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AAClB,CAAC,CAAC,CAAC"}

package/dist/scripts/pre-article-analysis/data-downloader.d.ts

@@ -0,0 +1,80 @@
+/**
+ * @module pre-article-analysis/data-downloader
+ * @description Downloads all relevant parliamentary documents from riksdag-regering-mcp
+ * for the current Swedish parliamentary session (riksmöte). Returns typed `RawDocument[]`
+ * collections plus a manifest of which MCP tools returned successful results and how many
+ * documents each returned.
+ *
+ * Note: The download methods fetch session-wide latest documents (bounded by `limit`
+ * and `rm`). Date-specific filtering should be applied by the caller after download
+ * (e.g., filtering by the `datum` field on each `RawDocument`).
+ *
+ * This module is intentionally side-effect-free with respect to the filesystem;
+ * callers are responsible for writing any output.
+ *
+ * @author Hack23 AB
+ * @license Apache-2.0
+ */
+import type { RawDocument } from '../data-transformers/types.js';
+import type { MCPClient } from '../mcp-client/client.js';
+/** Supported document type keys for scoped downloads. */
+export type DocumentTypeKey = 'propositions' | 'motions' | 'committeeReports' | 'votes' | 'speeches' | 'questions' | 'interpellations';
+export interface DownloadedData {
+    /** Government propositions */
+    propositions: RawDocument[];
+    /** Parliamentary motions */
+    motions: RawDocument[];
+    /** Committee reports */
+    committeeReports: RawDocument[];
+    /** Parliamentary votes */
+    votes: RawDocument[];
+    /** Parliamentary speeches (anföranden) */
+    speeches: RawDocument[];
+    /** Written questions */
+    questions: RawDocument[];
+    /** Interpellations */
+    interpellations: RawDocument[];
+}
+export interface DownloadManifest {
+    /** Names of MCP tools with successful fetch + post-processing during download */
+    dataSources: string[];
+    /** Document counts per type */
+    docCounts: Record<string, number>;
+    /** Total download duration in milliseconds */
+    durationMs: number;
+}
+export interface DownloadResult {
+    data: DownloadedData;
+    manifest: DownloadManifest;
+}
+/**
+ * Download all relevant documents from riksdag-regering-mcp.
+ *
+ * Each document type is fetched with a configurable limit (default 20).
+ * Failures in individual calls are caught so a partial result is returned
+ * rather than failing the entire pipeline.
+ *
+ * When `docTypes` is provided, only the listed document types are fetched.
+ * This prevents multiple workflows (e.g. propositions and committee-reports)
+ * from downloading the same documents and writing conflicting analysis.
+ *
+ * @param client   - MCPClient instance (caller-supplied for testability)
+ * @param options  - Optional overrides for limits, riksmöte, and document type scoping
+ */
+export declare function downloadAllDocuments(client: MCPClient, options?: {
+    limit?: number;
+    rm?: string;
+    docTypes?: DocumentTypeKey[];
+}): Promise<DownloadResult>;
+/**
+ * Flatten all downloaded document collections into a single array,
+ * deduplicated by a best-effort document identifier.
+ *
+ * The primary deduplication key is `dok_id` when present. If `dok_id` is
+ * missing or empty, the function falls back in order to:
+ * `dokument_id`, `id`, `dok_url`, `url`, `rel_dok_id`, `titel`, and `title`.
+ * The first non-empty string among these fields is used as the dedup key,
+ * and documents resolving to the same key are treated as duplicates.
+ */
+export declare function flattenDocuments(data: DownloadedData): RawDocument[];
+//# sourceMappingURL=data-downloader.d.ts.map

package/dist/scripts/pre-article-analysis/data-downloader.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"data-downloader.d.ts","sourceRoot":"","sources":["../../../scripts/pre-article-analysis/data-downloader.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AAEH,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,+BAA+B,CAAC;AACjE,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,yBAAyB,CAAC;AAMzD,yDAAyD;AACzD,MAAM,MAAM,eAAe,GACvB,cAAc,GACd,SAAS,GACT,kBAAkB,GAClB,OAAO,GACP,UAAU,GACV,WAAW,GACX,iBAAiB,CAAC;AAEtB,MAAM,WAAW,cAAc;IAC7B,8BAA8B;IAC9B,YAAY,EAAE,WAAW,EAAE,CAAC;IAC5B,4BAA4B;IAC5B,OAAO,EAAE,WAAW,EAAE,CAAC;IACvB,wBAAwB;IACxB,gBAAgB,EAAE,WAAW,EAAE,CAAC;IAChC,0BAA0B;IAC1B,KAAK,EAAE,WAAW,EAAE,CAAC;IACrB,0CAA0C;IAC1C,QAAQ,EAAE,WAAW,EAAE,CAAC;IACxB,wBAAwB;IACxB,SAAS,EAAE,WAAW,EAAE,CAAC;IACzB,sBAAsB;IACtB,eAAe,EAAE,WAAW,EAAE,CAAC;CAChC;AAED,MAAM,WAAW,gBAAgB;IAC/B,iFAAiF;IACjF,WAAW,EAAE,MAAM,EAAE,CAAC;IACtB,+BAA+B;IAC/B,SAAS,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAClC,8CAA8C;IAC9C,UAAU,EAAE,MAAM,CAAC;CACpB;AAED,MAAM,WAAW,cAAc;IAC7B,IAAI,EAAE,cAAc,CAAC;IACrB,QAAQ,EAAE,gBAAgB,CAAC;CAC5B;AA8DD;;;;;;;;;;;;;GAaG;AACH,wBAAsB,oBAAoB,CACxC,MAAM,EAAE,SAAS,EACjB,OAAO,GAAE;IAAE,KAAK,CAAC,EAAE,MAAM,CAAC;IAAC,EAAE,CAAC,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,EAAE,eAAe,EAAE,CAAA;CAAO,GAC1E,OAAO,CAAC,cAAc,CAAC,CAqHzB;AAED;;;;;;;;;GASG;AACH,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,cAAc,GAAG,WAAW,EAAE,CAgCpE"}

package/dist/scripts/pre-article-analysis/data-downloader.js

@@ -0,0 +1,230 @@
+/**
+ * @module pre-article-analysis/data-downloader
+ * @description Downloads all relevant parliamentary documents from riksdag-regering-mcp
+ * for the current Swedish parliamentary session (riksmöte). Returns typed `RawDocument[]`
+ * collections plus a manifest of which MCP tools returned successful results and how many
+ * documents each returned.
+ *
+ * Note: The download methods fetch session-wide latest documents (bounded by `limit`
+ * and `rm`). Date-specific filtering should be applied by the caller after download
+ * (e.g., filtering by the `datum` field on each `RawDocument`).
+ *
+ * This module is intentionally side-effect-free with respect to the filesystem;
+ * callers are responsible for writing any output.
+ *
+ * @author Hack23 AB
+ * @license Apache-2.0
+ */
+// ---------------------------------------------------------------------------
+// Helpers
+// ---------------------------------------------------------------------------
+function normalise(raw) {
+    return raw.filter(Boolean);
+}
+/** All internal fetch task names, kept in sync with the `fetchTasks` array
+ *  inside `downloadAllDocuments()`.  Used to derive the `FetchTaskName` type
+ *  and to validate the `FETCH_TASK_TYPE_MAP` at compile time. */
+const FETCH_TASK_NAMES = [
+    'fetchPropositions',
+    'fetchMotions',
+    'fetchCommitteeReports',
+    'fetchVotingRecords',
+    'searchSpeeches',
+    'fetchWrittenQuestions',
+    'fetchInterpellations',
+];
+/** Maps internal fetch task names to their corresponding DocumentTypeKey.
+ *  `satisfies` ensures every FetchTaskName maps to a valid DocumentTypeKey
+ *  at compile time — adding/renaming a task without updating the map is a
+ *  compile error. */
+const FETCH_TASK_TYPE_MAP = {
+    fetchPropositions: 'propositions',
+    fetchMotions: 'motions',
+    fetchCommitteeReports: 'committeeReports',
+    fetchVotingRecords: 'votes',
+    searchSpeeches: 'speeches',
+    fetchWrittenQuestions: 'questions',
+    fetchInterpellations: 'interpellations',
+};
+/**
+ * Returns the current Swedish parliamentary session (riksmöte) in `YYYY/YY` format.
+ * The Swedish parliamentary year runs from October to September:
+ *   - October–December of year N: session is `N/N+1`
+ *   - January–September of year N: session is `N-1/N`
+ *
+ * Examples: 2025-11 → "2025/26", 2026-03 → "2025/26", 2026-10 → "2026/27"
+ */
+function currentRm() {
+    const now = new Date();
+    const year = now.getUTCFullYear();
+    const month = now.getUTCMonth() + 1; // 1-based
+    // Swedish parliamentary year runs roughly October–September
+    if (month >= 10) {
+        return `${year}/${String(year + 1).slice(-2)}`;
+    }
+    return `${year - 1}/${String(year).slice(-2)}`;
+}
+// ---------------------------------------------------------------------------
+// Download orchestration
+// ---------------------------------------------------------------------------
+/**
+ * Download all relevant documents from riksdag-regering-mcp.
+ *
+ * Each document type is fetched with a configurable limit (default 20).
+ * Failures in individual calls are caught so a partial result is returned
+ * rather than failing the entire pipeline.
+ *
+ * When `docTypes` is provided, only the listed document types are fetched.
+ * This prevents multiple workflows (e.g. propositions and committee-reports)
+ * from downloading the same documents and writing conflicting analysis.
+ *
+ * @param client   - MCPClient instance (caller-supplied for testability)
+ * @param options  - Optional overrides for limits, riksmöte, and document type scoping
+ */
+export async function downloadAllDocuments(client, options = {}) {
+    const start = Date.now();
+    const limit = options.limit ?? 20;
+    const rm = options.rm ?? currentRm();
+    const docTypes = options.docTypes ?? null; // null means fetch all types
+    const dataSources = [];
+    const data = {
+        propositions: [],
+        motions: [],
+        committeeReports: [],
+        votes: [],
+        speeches: [],
+        questions: [],
+        interpellations: [],
+    };
+    // Run independent MCP fetches in parallel to reduce total latency while
+    // still collecting partial results on failure.
+    const fetchTasks = [
+        {
+            name: 'fetchPropositions',
+            source: 'get_propositioner',
+            fetch: () => client.fetchPropositions(limit, rm),
+            assign: (raw) => { data.propositions = raw; },
+        },
+        {
+            name: 'fetchMotions',
+            source: 'get_motioner',
+            fetch: () => client.fetchMotions(limit, rm),
+            assign: (raw) => { data.motions = raw; },
+        },
+        {
+            name: 'fetchCommitteeReports',
+            source: 'get_betankanden',
+            fetch: () => client.fetchCommitteeReports(limit, rm),
+            assign: (raw) => { data.committeeReports = raw; },
+        },
+        {
+            name: 'fetchVotingRecords',
+            source: 'search_voteringar',
+            fetch: () => client.fetchVotingRecords({ limit, rm }),
+            assign: (raw) => { data.votes = raw; },
+        },
+        {
+            name: 'searchSpeeches',
+            source: 'search_anforanden',
+            fetch: () => client.searchSpeeches({ limit, rm }),
+            assign: (raw) => { data.speeches = raw; },
+        },
+        {
+            name: 'fetchWrittenQuestions',
+            source: 'get_fragor',
+            fetch: () => client.fetchWrittenQuestions({ limit, rm }),
+            assign: (raw) => { data.questions = raw; },
+        },
+        {
+            name: 'fetchInterpellations',
+            source: 'get_interpellationer',
+            fetch: () => client.fetchInterpellations({ limit, rm }),
+            assign: (raw) => { data.interpellations = raw; },
+        },
+    ];
+    // When docTypes is specified, only fetch the listed document types.
+    // task.name is typed as FetchTaskName via the `as const` assertion on
+    // fetchTasks, so the lookup is safe without a cast.
+    const activeTasks = docTypes
+        ? fetchTasks.filter(task => {
+            const mapped = FETCH_TASK_TYPE_MAP[task.name];
+            return docTypes.includes(mapped);
+        })
+        : fetchTasks;
+    const fetchResults = await Promise.allSettled(activeTasks.map(task => task.fetch()));
+    fetchResults.forEach((result, index) => {
+        const task = activeTasks[index];
+        if (result.status === 'fulfilled') {
+            try {
+                task.assign(normalise(result.value));
+                dataSources.push(task.source);
+            }
+            catch (err) {
+                console.warn(`[pre-analysis] ${task.name} post-processing failed:`, err instanceof Error ? err.message : String(err));
+            }
+        }
+        else {
+            console.warn(`[pre-analysis] ${task.name} failed:`, result.reason instanceof Error ? result.reason.message : String(result.reason));
+        }
+    });
+    const docCounts = {
+        propositions: data.propositions.length,
+        motions: data.motions.length,
+        committeeReports: data.committeeReports.length,
+        votes: data.votes.length,
+        speeches: data.speeches.length,
+        questions: data.questions.length,
+        interpellations: data.interpellations.length,
+    };
+    return {
+        data,
+        manifest: {
+            dataSources,
+            docCounts,
+            durationMs: Date.now() - start,
+        },
+    };
+}
+/**
+ * Flatten all downloaded document collections into a single array,
+ * deduplicated by a best-effort document identifier.
+ *
+ * The primary deduplication key is `dok_id` when present. If `dok_id` is
+ * missing or empty, the function falls back in order to:
+ * `dokument_id`, `id`, `dok_url`, `url`, `rel_dok_id`, `titel`, and `title`.
+ * The first non-empty string among these fields is used as the dedup key,
+ * and documents resolving to the same key are treated as duplicates.
+ */
+export function flattenDocuments(data) {
+    const all = [
+        ...data.propositions,
+        ...data.motions,
+        ...data.committeeReports,
+        ...data.votes,
+        ...data.speeches,
+        ...data.questions,
+        ...data.interpellations,
+    ];
+    const seen = new Set();
+    return all.filter(doc => {
+        if (!doc)
+            return false;
+        const record = doc;
+        const idCandidates = [
+            record['dok_id'],
+            record['dokument_id'],
+            record['id'],
+            record['dok_url'],
+            record['url'],
+            record['rel_dok_id'],
+            record['titel'],
+            record['title'],
+        ];
+        const id = idCandidates.find((candidate) => typeof candidate === 'string' && candidate.trim().length > 0)?.trim() ?? '';
+        if (!id || seen.has(id))
+            return false;
+        seen.add(id);
+        return true;
+    });
+}
+//# sourceMappingURL=data-downloader.js.map

package/dist/scripts/pre-article-analysis/data-downloader.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"data-downloader.js","sourceRoot":"","sources":["../../../scripts/pre-article-analysis/data-downloader.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AAkDH,8EAA8E;AAC9E,UAAU;AACV,8EAA8E;AAE9E,SAAS,SAAS,CAAC,GAAc;IAC/B,OAAQ,GAAqB,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;AAChD,CAAC;AAED;;iEAEiE;AACjE,MAAM,gBAAgB,GAAG;IACvB,mBAAmB;IACnB,cAAc;IACd,uBAAuB;IACvB,oBAAoB;IACpB,gBAAgB;IAChB,uBAAuB;IACvB,sBAAsB;CACd,CAAC;AAIX;;;qBAGqB;AACrB,MAAM,mBAAmB,GAA2C;IAClE,iBAAiB,EAAE,cAAc;IACjC,YAAY,EAAE,SAAS;IACvB,qBAAqB,EAAE,kBAAkB;IACzC,kBAAkB,EAAE,OAAO;IAC3B,cAAc,EAAE,UAAU;IAC1B,qBAAqB,EAAE,WAAW;IAClC,oBAAoB,EAAE,iBAAiB;CACkB,CAAC;AAE5D;;;;;;;GAOG;AACH,SAAS,SAAS;IAChB,MAAM,GAAG,GAAG,IAAI,IAAI,EAAE,CAAC;IACvB,MAAM,IAAI,GAAG,GAAG,CAAC,cAAc,EAAE,CAAC;IAClC,MAAM,KAAK,GAAG,GAAG,CAAC,WAAW,EAAE,GAAG,CAAC,CAAC,CAAC,UAAU;IAC/C,4DAA4D;IAC5D,IAAI,KAAK,IAAI,EAAE,EAAE,CAAC;QAChB,OAAO,GAAG,IAAI,IAAI,MAAM,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IACjD,CAAC;IACD,OAAO,GAAG,IAAI,GAAG,CAAC,IAAI,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;AACjD,CAAC;AAED,8EAA8E;AAC9E,yBAAyB;AACzB,8EAA8E;AAE9E;;;;;;;;;;;;;GAaG;AACH,MAAM,CAAC,KAAK,UAAU,oBAAoB,CACxC,MAAiB,EACjB,UAAyE,EAAE;IAE3E,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IACzB,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,IAAI,EAAE,CAAC;IAClC,MAAM,EAAE,GAAG,OAAO,CAAC,EAAE,IAAI,SAAS,EAAE,CAAC;IACrC,MAAM,QAAQ,GAAG,OAAO,CAAC,QAAQ,IAAI,IAAI,CAAC,CAAC,6BAA6B;IAExE,MAAM,WAAW,GAAa,EAAE,CAAC;IACjC,MAAM,IAAI,GAAmB;QAC3B,YAAY,EAAE,EAAE;QAChB,OAAO,EAAE,EAAE;QACX,gBAAgB,EAAE,EAAE;QACpB,KAAK,EAAE,EAAE;QACT,QAAQ,EAAE,EAAE;QACZ,SAAS,EAAE,EAAE;QACb,eAAe,EAAE,EAAE;KACpB,CAAC;IAEF,wEAAwE;IACxE,+CAA+C;IAC/C,MAAM,UAAU,GAAG;QACjB;YACE,IAAI,EAAE,mBAAmB;YACzB,MAAM,EAAE,mBAAmB;YAC3B,KAAK,EAAE,GAAG,EAAE,CAAC,MAAM,CAAC,iBAAiB,CAAC,KAAK,EAAE,EAAE,CAAC;YAChD,MAAM,EAAE,CAAC,GAAkB,EAAE,EAAE,GAAG,IAAI,CAAC,YAAY,GAAG,GAAG,CAAC,CAAC,CAAC;SAC7D;QACD;YACE,IAAI,EAAE,cAAc;YACpB,MAAM,EAAE,cAAc;YACtB,KAAK,EAAE,GAAG,EAAE,CAAC,MAAM,CAAC,YAAY,CAAC,KAAK,EAAE,EAAE,CAAC;YAC3C,MAAM,EAAE,CAAC,GAAkB,EAAE,EAAE,GAAG,IAAI,CAAC,OAAO,GAAG,GAAG,CAAC,CAAC,CAAC;SACxD;QACD;YACE,IAAI,EAAE,uBAAuB;YAC7B,MAAM,EAAE,iBAAiB;YACzB,KAAK,EAAE,GAAG,EAAE,CAAC,MAAM,CAAC,qBAAqB,CAAC,KAAK,EAAE,EAAE,CAAC;YACpD,MAAM,EAAE,CAAC,GAAkB,EAAE,EAAE,GAAG,IAAI,CAAC,gBAAgB,GAAG,GAAG,CAAC,CAAC,CAAC;SACjE;QACD;YACE,IAAI,EAAE,oBAAoB;YAC1B,MAAM,EAAE,mBAAmB;YAC3B,KAAK,EAAE,GAAG,EAAE,CAAC,MAAM,CAAC,kBAAkB,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC;YACrD,MAAM,EAAE,CAAC,GAAkB,EAAE,EAAE,GAAG,IAAI,CAAC,KAAK,GAAG,GAAG,CAAC,CAAC,CAAC;SACtD;QACD;YACE,IAAI,EAAE,gBAAgB;YACtB,MAAM,EAAE,mBAAmB;YAC3B,KAAK,EAAE,GAAG,EAAE,CAAC,MAAM,CAAC,cAAc,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC;YACjD,MAAM,EAAE,CAAC,GAAkB,EAAE,EAAE,GAAG,IAAI,CAAC,QAAQ,GAAG,GAAG,CAAC,CAAC,CAAC;SACzD;QACD;YACE,IAAI,EAAE,uBAAuB;YAC7B,MAAM,EAAE,YAAY;YACpB,KAAK,EAAE,GAAG,EAAE,CAAC,MAAM,CAAC,qBAAqB,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC;YACxD,MAAM,EAAE,CAAC,GAAkB,EAAE,EAAE,GAAG,IAAI,CAAC,SAAS,GAAG,GAAG,CAAC,CAAC,CAAC;SAC1D;QACD;YACE,IAAI,EAAE,sBAAsB;YAC5B,MAAM,EAAE,sBAAsB;YAC9B,KAAK,EAAE,GAAG,EAAE,CAAC,MAAM,CAAC,oBAAoB,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC;YACvD,MAAM,EAAE,CAAC,GAAkB,EAAE,EAAE,GAAG,IAAI,CAAC,eAAe,GAAG,GAAG,CAAC,CAAC,CAAC;SAChE;KACO,CAAC;IAEX,oEAAoE;IACpE,sEAAsE;IACtE,oDAAoD;IACpD,MAAM,WAAW,GAAG,QAAQ;QAC1B,CAAC,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE;YACvB,MAAM,MAAM,GAAG,mBAAmB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAC9C,OAAO,QAAQ,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;QACnC,CAAC,CAAC;QACJ,CAAC,CAAC,UAAU,CAAC;IAEf,MAAM,YAAY,GAAG,MAAM,OAAO,CAAC,UAAU,CAC3C,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC,CACtC,CAAC;IAEF,YAAY,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE;QACrC,MAAM,IAAI,GAAG,WAAW,CAAC,KAAK,CAAE,CAAC;QAEjC,IAAI,MAAM,CAAC,MAAM,KAAK,WAAW,EAAE,CAAC;YAClC,IAAI,CAAC;gBACH,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,MAAM,CAAC,KAAkB,CAAC,CAAC,CAAC;gBAClD,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YAChC,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,OAAO,CAAC,IAAI,CACV,kBAAkB,IAAI,CAAC,IAAI,0BAA0B,EACrD,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CACjD,CAAC;YACJ,CAAC;QACH,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,IAAI,CACV,kBAAkB,IAAI,CAAC,IAAI,UAAU,EACrC,MAAM,CAAC,MAAM,YAAY,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAC/E,CAAC;QACJ,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,MAAM,SAAS,GAA2B;QACxC,YAAY,EAAE,IAAI,CAAC,YAAY,CAAC,MAAM;QACtC,OAAO,EAAE,IAAI,CAAC,OAAO,CAAC,MAAM;QAC5B,gBAAgB,EAAE,IAAI,CAAC,gBAAgB,CAAC,MAAM;QAC9C,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,MAAM;QACxB,QAAQ,EAAE,IAAI,CAAC,QAAQ,CAAC,MAAM;QAC9B,SAAS,EAAE,IAAI,CAAC,SAAS,CAAC,MAAM;QAChC,eAAe,EAAE,IAAI,CAAC,eAAe,CAAC,MAAM;KAC7C,CAAC;IAEF,OAAO;QACL,IAAI;QACJ,QAAQ,EAAE;YACR,WAAW;YACX,SAAS;YACT,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,KAAK;SAC/B;KACF,CAAC;AACJ,CAAC;AAED;;;;;;;;;GASG;AACH,MAAM,UAAU,gBAAgB,CAAC,IAAoB;IACnD,MAAM,GAAG,GAAkB;QACzB,GAAG,IAAI,CAAC,YAAY;QACpB,GAAG,IAAI,CAAC,OAAO;QACf,GAAG,IAAI,CAAC,gBAAgB;QACxB,GAAG,IAAI,CAAC,KAAK;QACb,GAAG,IAAI,CAAC,QAAQ;QAChB,GAAG,IAAI,CAAC,SAAS;QACjB,GAAG,IAAI,CAAC,eAAe;KACxB,CAAC;IAEF,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;IAC/B,OAAO,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE;QACtB,IAAI,CAAC,GAAG;YAAE,OAAO,KAAK,CAAC;QACvB,MAAM,MAAM,GAAG,GAA8B,CAAC;QAC9C,MAAM,YAAY,GAAG;YACnB,MAAM,CAAC,QAAQ,CAAC;YAChB,MAAM,CAAC,aAAa,CAAC;YACrB,MAAM,CAAC,IAAI,CAAC;YACZ,MAAM,CAAC,SAAS,CAAC;YACjB,MAAM,CAAC,KAAK,CAAC;YACb,MAAM,CAAC,YAAY,CAAC;YACpB,MAAM,CAAC,OAAO,CAAC;YACf,MAAM,CAAC,OAAO,CAAC;SAChB,CAAC;QACF,MAAM,EAAE,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC,SAAS,EAAuB,EAAE,CAC9D,OAAO,SAAS,KAAK,QAAQ,IAAI,SAAS,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAC7D,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;QAChB,IAAI,CAAC,EAAE,IAAI,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;YAAE,OAAO,KAAK,CAAC;QACtC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACb,OAAO,IAAI,CAAC;IACd,CAAC,CAAC,CAAC;AACL,CAAC"}