@koda-sl/baker-cli 0.574.0 → 0.575.0-dev.85f21462e

package/dist/top-companies-RC5KTWQ5.js

@@ -0,0 +1,246 @@
+import {
+  companyNameKey
+} from "./chunk-AYKV2B6X.js";
+import {
+  knowledgeTargetsHint,
+  targetsArg
+} from "./chunk-UZ5ZC43B.js";
+import {
+  targetsHints
+} from "./chunk-IR2NRXHA.js";
+import "./chunk-X7NRAGJ5.js";
+import "./chunk-GLY3CB2F.js";
+import "./chunk-IPEQHZ5H.js";
+import "./chunk-GRQ2AXDX.js";
+import {
+  csvOrJson,
+  daysAgoIso,
+  handleLinkedinError,
+  optionalAccountIdArg,
+  todayIso
+} from "./chunk-LQRPFL2N.js";
+import {
+  TABLE_OUTPUT_FORMATS,
+  outputFormatRefusal,
+  writeAdsJson,
+  writeAdsOutput,
+  writeJsonEnvelope
+} from "./chunk-VJUVTMHY.js";
+import {
+  NO_CACHE_ARG
+} from "./chunk-HPF3USL2.js";
+import {
+  organizationUrnOf,
+  targetsSummary
+} from "./chunk-BTOGDXA4.js";
+import "./chunk-QV6AST5T.js";
+import "./chunk-WCU362Z6.js";
+import "./chunk-GIWHHCD6.js";
+import "./chunk-SOPVIIUW.js";
+import "./chunk-G4FG5EQO.js";
+import "./chunk-3WOFZTGC.js";
+import "./chunk-IMS53EQU.js";
+import {
+  apiPost
+} from "./chunk-JCFZ7ECN.js";
+import "./chunk-OFSUVUXT.js";
+import "./chunk-OMLZ35HJ.js";
+import "./chunk-NXNOXZGB.js";
+import "./chunk-YQSBHWOI.js";
+
+// ../api/src/accounts/targetMatch.ts
+function indexTargets(targets) {
+  const index = { byOrg: /* @__PURE__ */ new Map(), byDomain: /* @__PURE__ */ new Map(), byName: /* @__PURE__ */ new Map() };
+  const claim = (map, key, target) => {
+    if (key && !map.has(key)) map.set(key, target);
+  };
+  for (const target of targets) {
+    claim(index.byOrg, target.organizationUrn ? organizationUrnOf(target.organizationUrn) : void 0, target);
+    claim(index.byDomain, target.domain, target);
+    claim(index.byName, target.name ? companyNameKey(target.name) : void 0, target);
+  }
+  return index;
+}
+function targetOf(company, index) {
+  const byOrg = index.byOrg.get(company.org);
+  if (byOrg) return { target: byOrg, by: "organization" };
+  const byDomain = company.domain ? index.byDomain.get(company.domain) : void 0;
+  if (byDomain) return { target: byDomain, by: "domain" };
+  const key = company.name ? companyNameKey(company.name) : "";
+  const byName = key !== "" ? index.byName.get(key) : void 0;
+  return byName ? { target: byName, by: "name" } : null;
+}
+function matchTargets(companies, targets) {
+  const index = indexTargets(targets);
+  const matched = companies.flatMap((company) => {
+    const found = targetOf(company, index);
+    return found ? [{ company, ...found }] : [];
+  });
+  const reached = new Set(matched.map(({ target }) => target.domain));
+  return { matched, notMatched: targets.filter((target) => !reached.has(target.domain)) };
+}
+
+// src/commands/ads/linkedin/top-companies.ts
+import { defineCommand } from "citty";
+var COMPANIES_HINT = "For whole-account reach over months (one row per company, --by month|campaign, --with-domains), use `baker ads linkedin companies`.";
+function topTargetCompanies(rows, targets) {
+  const companies = rows.flatMap((row) => {
+    const org = row.pivot.find((entry) => entry.urn.startsWith("urn:li:organization:"));
+    return org ? [{ row, org: org.urn, ...org.name ? { name: org.name } : {} }] : [];
+  });
+  const { matched } = matchTargets(companies, targets);
+  const kept = matched.map(({ company, target, by }) => ({ ...company.row, target: { ...target, matchedBy: by } }));
+  const reached = kept.map((row) => ({ domain: row.target.domain, engagements: Number(row.totalEngagements ?? 0) }));
+  return { rows: kept, targets: targetsSummary(targets, reached) };
+}
+var ENGAGEMENT_INTENTS = /* @__PURE__ */ new Set(["engagement"]);
+function topCompaniesIntent(intent, withTargets) {
+  if (!withTargets) return { intent: intent ?? "baseline" };
+  if (intent === void 0 || ENGAGEMENT_INTENTS.has(intent)) return { intent: intent ?? "engagement" };
+  return {
+    invalid: `--targets counts which target accounts engaged, and --intent ${intent} reads no engagements. Drop --intent, or pass --intent engagement.`
+  };
+}
+var TOP_COMPANIES_ROWS = 500;
+function topTargetsHint(summary, rowsRead) {
+  const nameless = summary.notReached.filter((target) => !target.name && !target.organizationUrn).length;
+  const full = rowsRead >= TOP_COMPANIES_ROWS;
+  if (summary.notReached.length === 0 || nameless === 0 && !full) return null;
+  const reasons = [
+    ...nameless > 0 ? [
+      `${nameless === 1 ? "1 target account has" : `${nameless} target accounts have`} neither a name nor a LinkedIn page, and this read matches on nothing else`
+    ] : [],
+    ...full ? [`this read covers only the ${TOP_COMPANIES_ROWS} companies with the most impressions`] : []
+  ];
+  return `Some targets never reached may have been: ${reasons.join("; ")}. \`baker ads linkedin companies --targets\` matches websites too, across every company reached.`;
+}
+function topCompaniesRead(args) {
+  const formatRefusal = outputFormatRefusal(csvOrJson(args), TABLE_OUTPUT_FORMATS);
+  if (formatRefusal !== null) return { invalid: formatRefusal };
+  let targets;
+  try {
+    targets = targetsArg(args.targets, args.tier);
+  } catch (err) {
+    return { invalid: `--targets: ${err instanceof Error ? err.message : String(err)}` };
+  }
+  const ids = String(args["campaign-id"] ?? "").split(",").map((id) => id.trim()).filter(Boolean);
+  if (ids.length > 0) return { level: "campaign", scope: ids, targets };
+  const accountId = optionalAccountIdArg(args);
+  if (accountId === void 0) {
+    return {
+      invalid: "This company has several LinkedIn ad accounts selected: pass --account-id for a whole-account read, or --campaign-id. Run `baker ads linkedin accounts` to list them."
+    };
+  }
+  return { level: "account", scope: [accountId], targets };
+}
+function topCompaniesAnswer(rows, targets, topN) {
+  const sorted = [...rows].sort((a, b) => Number(b.impressions ?? 0) - Number(a.impressions ?? 0));
+  const narrowed = targets ? topTargetCompanies(sorted, targets.targets) : null;
+  const top = narrowed ? narrowed.rows : sorted.slice(0, topN);
+  const missedHint = narrowed ? topTargetsHint(narrowed.targets, rows.length) : null;
+  const hints = [
+    ...targets ? targetsHints(narrowed?.targets, targets.unreadable) : knowledgeTargetsHint(),
+    ...missedHint ? [missedHint] : [],
+    COMPANIES_HINT
+  ];
+  return { top, narrowed, hints };
+}
+var topCompaniesCommand = defineCommand({
+  meta: {
+    name: "top-companies",
+    description: `Top companies whose employees saw / clicked / converted on some campaigns, or the whole ad account.
+
+Wrapper for: analytics --pivot company. Returns rows sorted by impressions desc.
+Each row's pivot[] contains { urn, id, name } \u2014 org names are resolved automatically.
+Scoring a target-account list? Pass --targets (one call; every target's row, plus the ones never reached) \u2014
+never --top-n 500 joined by hand. For months of whole-account reach, use \`baker ads linkedin companies\`.
+
+Demographic data is delayed 12-24h with a \u22653-event privacy floor.
+
+Examples:
+  baker ads linkedin top-companies --campaign-id 1234 --last-days 30
+  baker ads linkedin top-companies --campaign-id 1,2,3 --intent revenue --last-days 90 --top-n 50
+  baker ads linkedin top-companies --campaign-id 1234 --output csv
+  baker ads linkedin top-companies --targets targets.csv --tier 1 --intent engagement --last-days 60   # whole account`
+  },
+  args: {
+    "campaign-id": { type: "string", description: "Comma-separated campaign IDs (default: the whole ad account)" },
+    "account-id": {
+      type: "string",
+      description: "Ad account (numeric or urn:li:sponsoredAccount:N) for a whole-account read. Default: the only selected one"
+    },
+    targets: {
+      type: "string",
+      description: "Only these target accounts: the file `node __tooling__/scripts/target-accounts.mjs --out targets.csv` writes (or any .csv/.json `baker accounts --targets` takes). Matched on LinkedIn page, then name; --top-n does not apply"
+    },
+    tier: { type: "string", description: "With --targets: only this tier's accounts" },
+    intent: {
+      type: "string",
+      description: "baseline | revenue | funnel | engagement | lead-gen (default: baseline; engagement with --targets)"
+    },
+    "top-n": { type: "string", description: "Top rows by impressions (default: 25)" },
+    start: { type: "string", description: "Start date YYYY-MM-DD" },
+    end: { type: "string", description: "End date YYYY-MM-DD" },
+    "last-days": { type: "string", description: "Window relative to today (default: 30)" },
+    "no-cache": NO_CACHE_ARG,
+    output: { type: "string", description: "json|csv|jsonl|md", default: "json" }
+  },
+  run: async ({ args }) => {
+    const read = topCompaniesRead(args);
+    if ("invalid" in read) {
+      writeAdsJson({ ok: false, error: { code: "INVALID_ARGS", message: read.invalid } });
+      process.exit(1);
+    }
+    const { level, scope, targets } = read;
+    const chosen = topCompaniesIntent(args.intent, targets !== null);
+    if ("invalid" in chosen) {
+      writeAdsJson({ ok: false, error: { code: "INVALID_ARGS", message: chosen.invalid } });
+      process.exit(1);
+    }
+    const { intent } = chosen;
+    const topN = args["top-n"] ? Number(args["top-n"]) : 25;
+    const start = args.start ?? daysAgoIso(args["last-days"] ? Number(args["last-days"]) : 30);
+    const end = args.end ?? todayIso();
+    try {
+      const data = await apiPost("/api/ads/linkedin/analytics", {
+        request: {
+          level,
+          ids: scope,
+          intent,
+          pivot: "company",
+          start,
+          end,
+          granularity: "ALL",
+          limit: TOP_COMPANIES_ROWS
+        },
+        skipCache: Boolean(args["no-cache"])
+      });
+      const { top, narrowed, hints } = topCompaniesAnswer(data.rows, targets, topN);
+      const fmt = csvOrJson(args);
+      if (fmt !== "json") {
+        writeAdsOutput(top, fmt, targets ? [...data.fields, "target"] : data.fields, hints);
+        return;
+      }
+      writeJsonEnvelope({
+        ok: true,
+        data: {
+          query: { level, ids: scope, start, end, intent, ...targets ? {} : { topN } },
+          rows: top,
+          ...narrowed ? { targets: narrowed.targets } : {},
+          warnings: data.warnings
+        },
+        hints
+      });
+    } catch (err) {
+      handleLinkedinError(err);
+    }
+  }
+});
+export {
+  TOP_COMPANIES_ROWS,
+  topCompaniesCommand,
+  topCompaniesIntent,
+  topTargetCompanies,
+  topTargetsHint
+};
+//# sourceMappingURL=top-companies-RC5KTWQ5.js.map

package/dist/transcript-OLF62WCH.js

@@ -0,0 +1,165 @@
+import {
+  transcriptHints
+} from "./chunk-ALCRXFIZ.js";
+import {
+  handleResearchError,
+  writeResearchJson
+} from "./chunk-7PKEBTVP.js";
+import "./chunk-BTOGDXA4.js";
+import "./chunk-QV6AST5T.js";
+import "./chunk-WCU362Z6.js";
+import "./chunk-GIWHHCD6.js";
+import "./chunk-SOPVIIUW.js";
+import "./chunk-G4FG5EQO.js";
+import "./chunk-3WOFZTGC.js";
+import "./chunk-IMS53EQU.js";
+import {
+  apiPost
+} from "./chunk-JCFZ7ECN.js";
+import "./chunk-OFSUVUXT.js";
+import "./chunk-OMLZ35HJ.js";
+import {
+  registerSchema
+} from "./chunk-NXNOXZGB.js";
+import "./chunk-YQSBHWOI.js";
+
+// src/commands/research/transcript.ts
+import { mkdirSync, writeFileSync } from "fs";
+import { dirname, resolve } from "path";
+import { defineCommand } from "citty";
+var EXCERPT_CHARS = 3e3;
+var TIMEOUT_MS = 54e4;
+registerSchema({
+  command: "research.transcript",
+  description: "The words of one long recording \u2014 a keynote, earnings call, webinar or podcast \u2014 so a leader's own phrasing can feed 1:1 ad and page messaging. A YouTube link is read from its captions; any other link must be the audio or video file itself. Nothing is added to the video library. Start here: `baker research transcript <url> --out transcript.txt`.",
+  args: {
+    url: {
+      type: "string",
+      description: "A YouTube link, or a direct link to an audio/video file (.mp3, .mp4, .m4a, \u2026) (positional)",
+      required: true
+    },
+    full: {
+      type: "boolean",
+      description: `Return the whole text instead of the first ${EXCERPT_CHARS} characters`,
+      required: false,
+      default: false
+    },
+    out: {
+      type: "string",
+      description: "Write the whole text to this file and return its path instead of the text",
+      required: false
+    }
+  }
+});
+var FIELDS = {
+  provider: "youtube \u2014 the video's captions; deepgram \u2014 the recording transcribed",
+  title: "The video's title (YouTube only), or null",
+  duration_sec: "Length of the recording in seconds, or null when unknown",
+  language: "The language spoken, as the provider names it",
+  word_count: "Words in the whole recording",
+  text_chars: "Length of the whole text available on the server",
+  excerpt: "The opening of the text (default output)",
+  text: "The whole text (--full)",
+  file: "Where the whole text was written (--out)",
+  truncated: "The recording was longer than one transcript carries, so the end is missing"
+};
+function transcriptFix(code) {
+  if (code === "VALIDATION_ERROR" || code === "BAD_REQUEST") {
+    return {
+      action: "retry_with_changes",
+      explanation: "Pass a YouTube link, or the direct link to the audio or video file \u2014 for a podcast, the episode's download link (its RSS enclosure), not the episode page."
+    };
+  }
+  if (code === "CONFLICT") {
+    return {
+      action: "continue_without",
+      explanation: "There are no words to read in this recording. Look for another recording of the same person, or write the brief from what you have."
+    };
+  }
+  if (code === "RATE_LIMITED") {
+    return {
+      action: "continue_without",
+      explanation: "The transcriber is busy. Do the rest of the job first and try this recording once more later."
+    };
+  }
+  return {
+    action: "continue_without",
+    explanation: "This recording couldn't be transcribed. Don't loop on it \u2014 finish the brief from your other sources and name this recording as a gap."
+  };
+}
+var transcriptCommand = defineCommand({
+  meta: {
+    name: "transcript",
+    description: `The words of one long recording \u2014 a keynote, earnings call, webinar or podcast \u2014 so a leader's own phrasing can feed 1:1 ad and page messaging.
+
+A YouTube link is read from its captions. Any other link must be the audio or video file itself (.mp3, .mp4, .m4a, \u2026): for a podcast, the episode's download link, not its page. Nothing is added to the video library.
+
+Default output is compact \u2014 title, length, word count and the opening of the text. Use --out to write the whole text to a file and read the parts you need.
+
+Examples:
+  baker research transcript "https://www.youtube.com/watch?v=<id>"
+  baker research transcript "https://www.youtube.com/watch?v=<id>" --out keynote.txt
+  baker research transcript "https://traffic.megaphone.fm/<episode>.mp3" --full`
+  },
+  args: {
+    url: { type: "positional", description: "A YouTube link, or a direct audio/video file link", required: false },
+    full: { type: "boolean", description: "Return the whole text", required: false },
+    out: { type: "string", description: "Write the whole text to this file", required: false }
+  },
+  run: async ({ args }) => {
+    const url = args.url;
+    if (!url) {
+      writeResearchJson({
+        ok: false,
+        error: {
+          code: "VALIDATION_ERROR",
+          message: "A recording link is required",
+          fix: transcriptFix("VALIDATION_ERROR")
+        }
+      });
+      process.exit(1);
+    }
+    const body = { url };
+    try {
+      const transcript = await apiPost("/api/research/transcript", body, {
+        timeoutMs: TIMEOUT_MS
+      });
+      const out = args.out;
+      let delivered;
+      let excerpted = false;
+      if (out) {
+        const file = resolve(out);
+        mkdirSync(dirname(file), { recursive: true });
+        writeFileSync(file, transcript.text, "utf-8");
+        delivered = { file };
+      } else if (args.full) {
+        delivered = { text: transcript.text };
+      } else {
+        excerpted = transcript.text.length > EXCERPT_CHARS;
+        delivered = { excerpt: transcript.text.slice(0, EXCERPT_CHARS) };
+      }
+      writeResearchJson({
+        ok: true,
+        data: {
+          url: transcript.url,
+          provider: transcript.provider,
+          title: transcript.title,
+          duration_sec: transcript.durationSec,
+          language: transcript.language,
+          word_count: transcript.wordCount,
+          text_chars: transcript.textChars,
+          truncated: transcript.truncated,
+          ...delivered
+        },
+        fields: FIELDS,
+        hints: transcriptHints({ excerpted, truncated: transcript.truncated })
+      });
+    } catch (err) {
+      handleResearchError(err, transcriptFix);
+    }
+  }
+});
+export {
+  transcriptCommand
+};
+//# sourceMappingURL=transcript-OLF62WCH.js.map