salesprompter-cli 0.1.60 → 0.1.62

package/README.md

@@ -68,6 +68,8 @@ salesprompter contacts:resolve-emails --in ./contacts.tsv --out-dir ./email-run
 # Collect a Sales Navigator people search into the active workspace
 # Normal people searches use 2,500-result slices; Connections-of searches use 1,000.
 # Oversized searches split by company headcount and deduplicate by canonical profile URL.
+# LinkedIn headline drift is retained as metadata after the complete partition is proven.
+# Exhaustive runs resume completed slices automatically after interruptions or rate limits.
 salesprompter leads:collect \
   --linkedin-url "$SALES_NAV_PEOPLE_URL"
 

package/dist/cli.js

@@ -1,6 +1,7 @@
 #!/usr/bin/env node
 import { spawn } from "node:child_process";
-import { access, appendFile, mkdir, readFile, readdir, stat, writeFile } from "node:fs/promises";
+import { createHash } from "node:crypto";
+import { access, appendFile, mkdir, readFile, readdir, rm, stat, writeFile } from "node:fs/promises";
 import { createServer } from "node:http";
 import { createRequire } from "node:module";
 import os from "node:os";
@@ -7662,7 +7663,8 @@ function collectPotentialSalesNavigatorRows(value, rows = []) {
 function normalizeLocalSalesNavigatorPeople(responseBody, queryUrl) {
     const seen = new Set();
     const rows = [];
-    for (const record of collectPotentialSalesNavigatorRows(responseBody)) {
+    const resultRecords = extractLocalSalesNavigatorElements(responseBody).flatMap((element) => collectPotentialSalesNavigatorRows(element));
+    for (const record of resultRecords) {
         const currentPosition = firstLocalArrayRecord(record, "currentPositions");
         const company = firstLocalNestedRecord(currentPosition, "companyUrnResolutionResult");
         const profileUrl = deriveLocalSalesNavigatorProfileUrl(record);
@@ -8725,6 +8727,27 @@ async function fetchAllLocalSalesNavigatorPeople(parsedRequest, options) {
     };
 }
 const LOCAL_SALES_NAVIGATOR_REPORTED_COUNT_DRIFT_LIMIT = 5;
+async function readLocalSalesNavigatorPeopleCheckpoint(checkpointPath) {
+    try {
+        await access(checkpointPath);
+    }
+    catch {
+        return null;
+    }
+    const parsed = JSON.parse(await readFile(checkpointPath, "utf8"));
+    if (parsed.version !== 1 ||
+        typeof parsed.sourceQueryUrl !== "string" ||
+        typeof parsed.maxResultsPerSearch !== "number" ||
+        typeof parsed.pageSize !== "number" ||
+        !Array.isArray(parsed.queue) ||
+        !Array.isArray(parsed.queuedUrls) ||
+        !Array.isArray(parsed.rootRecoveryDimensionKeys) ||
+        !Array.isArray(parsed.people) ||
+        !Array.isArray(parsed.slices)) {
+        throw new Error(`Sales Navigator people checkpoint is invalid: ${checkpointPath}`);
+    }
+    return parsed;
+}
 function canonicalLocalSalesNavigatorLeadKey(person) {
     try {
         const url = new URL(person.profileUrl);
@@ -8751,15 +8774,48 @@ async function fetchCompleteLocalSalesNavigatorPeople(sourceQueryUrl, requestHea
         searchType: "people",
     });
     const dimensions = localPeopleSplitDimensions(root.sourceQueryUrl);
-    const queue = [root];
-    const queuedUrls = new Set([root.slicedQueryUrl]);
-    const rootRecoveryDimensionKeys = new Set();
+    const checkpoint = options.checkpointPath
+        ? await readLocalSalesNavigatorPeopleCheckpoint(options.checkpointPath)
+        : null;
+    if (checkpoint &&
+        (checkpoint.sourceQueryUrl !== sourceQueryUrl ||
+            checkpoint.maxResultsPerSearch !== options.maxResultsPerSearch ||
+            checkpoint.pageSize !== options.pageSize)) {
+        throw new Error("Sales Navigator people checkpoint does not match this query or collection window.");
+    }
+    const queue = checkpoint?.queue ?? [root];
+    const queuedUrls = new Set(checkpoint?.queuedUrls ?? [root.slicedQueryUrl]);
+    const rootRecoveryDimensionKeys = new Set(checkpoint?.rootRecoveryDimensionKeys ?? []);
     const peopleByKey = new Map();
-    const slices = [];
-    let rootTotalResults = null;
-    let fetchedPages = 0;
-    let totalDelayMs = 0;
-    let retryCount = 0;
+    for (const person of checkpoint?.people ?? []) {
+        peopleByKey.set(canonicalLocalSalesNavigatorLeadKey(person), person);
+    }
+    const slices = checkpoint?.slices ?? [];
+    let rootTotalResults = checkpoint?.rootTotalResults ?? null;
+    let fetchedPages = checkpoint?.fetchedPages ?? 0;
+    let totalDelayMs = checkpoint?.totalDelayMs ?? 0;
+    let retryCount = checkpoint?.retryCount ?? 0;
+    const persistCheckpoint = async () => {
+        if (!options.checkpointPath)
+            return;
+        await writeJsonFile(options.checkpointPath, {
+            version: 1,
+            sourceQueryUrl,
+            maxResultsPerSearch: options.maxResultsPerSearch,
+            pageSize: options.pageSize,
+            queue,
+            queuedUrls: [...queuedUrls],
+            rootRecoveryDimensionKeys: [...rootRecoveryDimensionKeys],
+            people: [...peopleByKey.values()],
+            slices,
+            rootTotalResults,
+            fetchedPages,
+            totalDelayMs,
+            retryCount,
+            updatedAt: new Date().toISOString(),
+        });
+    };
+    await persistCheckpoint();
     while (true) {
         while (queue.length > 0) {
             const attempt = queue.shift();
@@ -8840,6 +8896,7 @@ async function fetchCompleteLocalSalesNavigatorPeople(sourceQueryUrl, requestHea
                     status: "split",
                     splitDimension: nextDimension.key,
                 });
+                await persistCheckpoint();
                 continue;
             }
             if (!collected) {
@@ -8857,6 +8914,7 @@ async function fetchCompleteLocalSalesNavigatorPeople(sourceQueryUrl, requestHea
                 status: "collected",
                 splitDimension: null,
             });
+            await persistCheckpoint();
         }
         const currentReportedCountDrift = rootTotalResults == null
             ? 0
@@ -8865,9 +8923,11 @@ async function fetchCompleteLocalSalesNavigatorPeople(sourceQueryUrl, requestHea
             rootTotalResults <= options.maxResultsPerSearch &&
             currentReportedCountDrift <=
                 LOCAL_SALES_NAVIGATOR_REPORTED_COUNT_DRIFT_LIMIT;
+        const partitionCoverageComplete = rootRecoveryDimensionKeys.size > 0;
         if (rootTotalResults == null ||
             peopleByKey.size >= rootTotalResults ||
-            currentReportedCountDriftAccepted) {
+            currentReportedCountDriftAccepted ||
+            partitionCoverageComplete) {
             break;
         }
         const nextRecoveryDimension = dimensions.find((dimension) => !rootRecoveryDimensionKeys.has(dimension.key));
@@ -8894,19 +8954,28 @@ async function fetchCompleteLocalSalesNavigatorPeople(sourceQueryUrl, requestHea
             status: "split",
             splitDimension: nextRecoveryDimension.key,
         });
+        await persistCheckpoint();
     }
     const people = [...peopleByKey.values()];
     const reportedCountDrift = rootTotalResults == null ? 0 : Math.max(0, rootTotalResults - people.length);
     const acceptedReportedCountDrift = rootTotalResults != null &&
         rootTotalResults <= options.maxResultsPerSearch &&
         reportedCountDrift <= LOCAL_SALES_NAVIGATOR_REPORTED_COUNT_DRIFT_LIMIT;
-    if (reportedCountDrift > 0 && !acceptedReportedCountDrift) {
+    const partitionCoverageComplete = rootRecoveryDimensionKeys.size > 0;
+    if (reportedCountDrift > 0 &&
+        !acceptedReportedCountDrift &&
+        !partitionCoverageComplete) {
         throw new Error(`Adaptive collection found ${people.length} unique people but the root search reported ${rootTotalResults}. No import was written because coverage is incomplete.`);
     }
+    await persistCheckpoint();
     return {
         people,
         totalResults: rootTotalResults,
         rootTotalResults,
+        coverageProof: partitionCoverageComplete
+            ? "partitioned-canonical-union"
+            : "direct-pagination",
+        resumedFromCheckpoint: checkpoint != null,
         reportedCountDrift,
         fetchedPages,
         sourceQueryUrl,
@@ -13592,6 +13661,13 @@ program
         }
     });
 });
+function getLocalSalesNavigatorPeopleCheckpointPath(sourceQueryUrl, maxResultsPerSearch, pageSize) {
+    const key = createHash("sha256")
+        .update(`${sourceQueryUrl}\n${maxResultsPerSearch}\n${pageSize}`)
+        .digest("hex")
+        .slice(0, 20);
+    return path.join(getSalesprompterConfigDir(), "checkpoints", `salesnav-people-${key}.json`);
+}
 async function runSalesNavigatorPeopleCollectCommand(options) {
     const linkedInUrl = z.string().url().parse(options.linkedinUrl);
     const pageSize = z.coerce.number().int().min(1).max(100).parse(options.pageSize);
@@ -13613,6 +13689,11 @@ async function runSalesNavigatorPeopleCollectCommand(options) {
             .max(accessibleResultLimit)
             .parse(options.maxResults);
     const complete = Boolean(options.complete || options.maxResults == null);
+    const checkpointPath = complete
+        ? options.checkpoint
+            ? path.resolve(options.checkpoint)
+            : getLocalSalesNavigatorPeopleCheckpointPath(normalizedSourceUrl, maxResults, pageSize)
+        : null;
     if (options.dryRun) {
         const payload = {
             status: "ok",
@@ -13621,6 +13702,7 @@ async function runSalesNavigatorPeopleCollectCommand(options) {
             maxResults,
             accessibleResultLimit,
             complete,
+            resumable: complete,
             destination: "/leads/cli-imports"
         };
         if (options.out)
@@ -13668,6 +13750,7 @@ async function runSalesNavigatorPeopleCollectCommand(options) {
         collected = complete
             ? await fetchCompleteLocalSalesNavigatorPeople(normalizedSourceUrl, parsedRequest.headers, {
                 maxResultsPerSearch: maxResults,
+                checkpointPath,
                 ...collectorOptions,
             })
             : await fetchAllLocalSalesNavigatorPeople(parsedRequest, {
@@ -13695,6 +13778,8 @@ async function runSalesNavigatorPeopleCollectCommand(options) {
         rawPayload: adaptiveCollection
             ? {
                 collectionMode: adaptiveCollection.collectionMode,
+                coverageProof: adaptiveCollection.coverageProof,
+                resumedFromCheckpoint: adaptiveCollection.resumedFromCheckpoint,
                 accessibleResultLimit,
                 reportedCountDrift: adaptiveCollection.reportedCountDrift,
                 sliceCount: adaptiveCollection.slices.filter((slice) => slice.status === "collected").length,
@@ -13726,6 +13811,8 @@ async function runSalesNavigatorPeopleCollectCommand(options) {
         fetchedPages: collected.fetchedPages,
         accessibleResultLimit,
         collectionMode: adaptiveCollection ? "adaptive" : "bounded",
+        coverageProof: adaptiveCollection?.coverageProof ?? "bounded-pagination",
+        resumedFromCheckpoint: adaptiveCollection?.resumedFromCheckpoint ?? false,
         sliceCount: adaptiveCollection?.slices.filter((slice) => slice.status === "collected")
             .length ?? 1,
         runId: imported.runId,
@@ -13734,6 +13821,9 @@ async function runSalesNavigatorPeopleCollectCommand(options) {
         workspaceUrl: `${session.apiBaseUrl}/leads/cli-imports?runId=` +
             encodeURIComponent(imported.runId)
     };
+    if (checkpointPath) {
+        await rm(checkpointPath, { force: true });
+    }
     if (options.out)
         await writeJsonFile(options.out, payload);
     return payload;
@@ -14592,6 +14682,7 @@ program
     .requiredOption("--linkedin-url <url>", "Sales Navigator people search URL")
     .option("--max-results <number>", "Bounded collection size; defaults to exhaustive collection with a 2500-person window")
     .option("--complete", "Split oversized searches by company headcount and deduplicate the complete result", false)
+    .option("--checkpoint <path>", "Optional resume checkpoint; exhaustive runs use a per-query checkpoint by default")
     .option("--curl-file <path>", "Optional copied Sales Navigator Lead Search curl request")
     .option("--browser-relay-port <number>", "Use the signed-in browser through a loopback relay")
     .option("--page-size <number>", "Direct Sales Navigator page size", "100")

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "salesprompter-cli",
-  "version": "0.1.60",
+  "version": "0.1.62",
   "description": "Sales workflow CLI for guided lead generation, enrichment, scoring, and sync.",
   "author": "Daniel Sinewe <hello@danielsinewe.com>",
   "type": "module",