@aglyn/plugins-crm 1.0.0-beta.217 → 1.0.0-beta.219

package/src/lib/model/contact-remove.d.ts

@@ -0,0 +1,48 @@
+/**
+ * @license
+ * Copyright 2026 Aglyn LLC
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *   http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+/**
+ * The contract of `crm/contact-remove` (AGL-3338), client-safe: what the
+ * record page's Delete and the contacts bar's Remove send, and what each
+ * contact is answered with.
+ */
+/** The most contacts one request may name — the bar sends a selection in pieces this size. */
+export declare const CRM_CONTACT_REMOVE_MAX = 200;
+export interface ContactRemoveRequest {
+    /** The mounted site; at the organization level the record's own site, or nothing. */
+    hostId?: string | null;
+    /** The organization, which makes the call the route's organization variant. */
+    orgId?: string;
+    contactIds: string[];
+}
+/**
+ * One contact's answer: `detached` when other sites still hold the person
+ * and only this holder's half went, `deleted` when nobody else did and the
+ * document went with it — or refused, with a sentence written for the reader.
+ */
+export type ContactRemoveOutcome = {
+    contactId: string;
+    ok: true;
+    removed: 'detached' | 'deleted';
+} | {
+    contactId: string;
+    ok: false;
+    error: string;
+};
+export interface ContactRemoveResponse {
+    ok: true;
+    results: ContactRemoveOutcome[];
+}

package/src/lib/model/contact-remove.js

@@ -0,0 +1,22 @@
+/**
+ * @license
+ * Copyright 2026 Aglyn LLC
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *   http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */ /**
+ * The contract of `crm/contact-remove` (AGL-3338), client-safe: what the
+ * record page's Delete and the contacts bar's Remove send, and what each
+ * contact is answered with.
+ */ /** The most contacts one request may name — the bar sends a selection in pieces this size. */ export const CRM_CONTACT_REMOVE_MAX = 200;
+
+//# sourceMappingURL=contact-remove.js.map

package/src/lib/model/contact-remove.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../../../../../../../libs/plugins/crm/src/lib/model/contact-remove.ts"],"sourcesContent":["/**\n * @license\n * Copyright 2026 Aglyn LLC\n *\n * Licensed under the Apache License, Version 2.0 (the \"License\");\n * you may not use this file except in compliance with the License.\n * You may obtain a copy of the License at\n *\n *   http://www.apache.org/licenses/LICENSE-2.0\n *\n * Unless required by applicable law or agreed to in writing, software\n * distributed under the License is distributed on an \"AS IS\" BASIS,\n * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n * See the License for the specific language governing permissions and\n * limitations under the License.\n */\n\n/**\n * The contract of `crm/contact-remove` (AGL-3338), client-safe: what the\n * record page's Delete and the contacts bar's Remove send, and what each\n * contact is answered with.\n */\n\n/** The most contacts one request may name — the bar sends a selection in pieces this size. */\nexport const CRM_CONTACT_REMOVE_MAX = 200\n\nexport interface ContactRemoveRequest {\n  /** The mounted site; at the organization level the record's own site, or nothing. */\n  hostId?: string | null\n  /** The organization, which makes the call the route's organization variant. */\n  orgId?: string\n  contactIds: string[]\n}\n\n/**\n * One contact's answer: `detached` when other sites still hold the person\n * and only this holder's half went, `deleted` when nobody else did and the\n * document went with it — or refused, with a sentence written for the reader.\n */\nexport type ContactRemoveOutcome =\n  | { contactId: string; ok: true; removed: 'detached' | 'deleted' }\n  | { contactId: string; ok: false; error: string }\n\nexport interface ContactRemoveResponse {\n  ok: true\n  results: ContactRemoveOutcome[]\n}\n"],"names":["CRM_CONTACT_REMOVE_MAX"],"mappings":"AAAA;;;;;;;;;;;;;;;CAeC,GAED;;;;CAIC,GAED,4FAA4F,GAC5F,OAAO,MAAMA,yBAAyB,IAAG"}

package/src/lib/model/contacts-bulk-writes.d.ts

@@ -22,8 +22,7 @@
  * to an owner, file them under a company, move them along the funnel, let
  * them go. A contact's tags, owner and company are one holder's records in
  * that holder's facet, and a facet is the server's to write: the rules cannot
- * tell one field of it from another, so they leave a client nothing there
- * but letting a holder go. So every act but that one is a request to a route,
+ * tell one field of it from another. So every act is a request to a route,
  * and what this module decides is WHICH rows the request names — the bar
  * holds only the controls.
  *
@@ -37,23 +36,8 @@
  * "398 of 400" is a number nobody can act on; an address is. The route asks
  * the same questions of the stored document, so a row that changed since the
  * table read it is still refused by name rather than slipped past.
- *
- * ## Letting go is still written here
- *
- * Removing a contact from a site drops that holder's facet, consent entries,
- * capture attribution and scope tokens — the one change to a facet a holder
- * may make client-direct. Those patches are built here and applied in
- * batches with a per-row fallback, so a refused row is named rather than
- * failing the other 399 with it.
- *
- * Pure apart from the Firestore sentinels the detach carries: the runner takes
- * its writers as ports, so the bar wires `writeBatch`/`updateDoc`/`deleteDoc`
- * in and a spec wires a ledger in.
  */
 import { type ContactCompanyLinkState } from '@aglyn/aglyn';
-import { chunked } from './crm-bulk-writes';
-export declare const CONTACT_BULK_WRITE_CHUNK = 400;
-export { chunked };
 /** The record page's cap on a holder's tags — the same number, so the two agree. */
 export declare const CONTACT_TAGS_CAP = 20;
 /** As much of a table row as a bulk act reads. */
@@ -71,31 +55,16 @@ export interface ContactBulkRow {
     holderHostId?: string;
     /** THIS holder's tags, already read through the facet by the table. */
     tags?: string[];
-    /** The holder tokens — what `planContactDetach` counts. */
+    /** The holder tokens the row is shared with. */
     visibleTo?: string[];
     /** The link state the company planner reads — see `ContactRecord.companyLink`. */
     companyLink?: ContactCompanyLinkState;
 }
-/** One client-direct write to one contact document, named by the address it is about. */
-export type ContactBulkWrite = {
-    id: string;
-    email: string;
-} & ({
-    kind: 'update';
-    data: Record<string, unknown>;
-} | {
-    kind: 'delete';
-});
 /** A row an action deliberately left alone, and why. */
 export interface ContactBulkSkip {
     email: string;
     reason: string;
 }
-/** What a letting-go wants written, and what it declined to. */
-export interface ContactBulkPlan {
-    writes: ContactBulkWrite[];
-    skipped: ContactBulkSkip[];
-}
 /** The rows a request to a route names, and the rows it leaves out by name. */
 export interface ContactBulkSelection {
     rows: ContactBulkRow[];
@@ -133,41 +102,3 @@ export declare function planRemoveTag(rows: readonly ContactBulkRow[], tag: stri
  * request is not the place to guess.
  */
 export declare function planSetCompany(rows: readonly ContactBulkRow[], companyId: string | null): ContactBulkSelection;
-/**
- * Let every row go — the record page's DELETE IS A DETACH, per row.
- *
- * `planContactDetach` decides for each document whether this holder is the
- * last one (delete) or one of several (drop this group's facet, consent
- * entries, capture attribution and scope tokens). The `arrayRemove` and
- * `deleteField` sentinels are the ones the record page writes, so a row
- * detached here and a row detached there are the same document afterwards.
- *
- * ⛔ Not the erasure path: a privacy erasure removes the person everywhere
- * regardless of holders.
- */
-export declare function planDetach(rows: readonly ContactBulkRow[], group: {
-    groupId: string;
-    hostIds: readonly string[];
-}, nowMs: number): ContactBulkPlan;
-/** The writers the runner needs — Firestore's, or a spec's ledger. */
-export interface ContactBulkWriters {
-    /** Apply every write atomically, or throw. */
-    commitBatch: (writes: readonly ContactBulkWrite[]) => Promise<void>;
-    /** Apply one write, or throw. */
-    commitOne: (write: ContactBulkWrite) => Promise<void>;
-}
-export interface ContactBulkOutcome {
-    /** Rows written. */
-    done: number;
-    /** Rows the store refused, by address, with the reason it gave. */
-    refused: Array<{
-        email: string;
-        error: string;
-    }>;
-}
-/**
- * The shared runner over contact writes — batched, with a per-row pass for
- * the chunk that failed (`crm-bulk-writes.ts`) — reporting by ADDRESS, which
- * is the name a contacts report lists a row under.
- */
-export declare function runContactBulkWrites(writers: ContactBulkWriters, writes: readonly ContactBulkWrite[], chunkSize?: number): Promise<ContactBulkOutcome>;

package/src/lib/model/contacts-bulk-writes.js

@@ -1,4 +1,3 @@
-import { _ as _extends } from "@swc/helpers/_/_extends";
 /**
  * @license
  * Copyright 2026 Aglyn LLC
@@ -22,8 +21,7 @@ import { _ as _extends } from "@swc/helpers/_/_extends";
  * to an owner, file them under a company, move them along the funnel, let
  * them go. A contact's tags, owner and company are one holder's records in
  * that holder's facet, and a facet is the server's to write: the rules cannot
- * tell one field of it from another, so they leave a client nothing there
- * but letting a holder go. So every act but that one is a request to a route,
+ * tell one field of it from another. So every act is a request to a route,
  * and what this module decides is WHICH rows the request names — the bar
  * holds only the controls.
  *
@@ -37,27 +35,7 @@ import { _ as _extends } from "@swc/helpers/_/_extends";
  * "398 of 400" is a number nobody can act on; an address is. The route asks
  * the same questions of the stored document, so a row that changed since the
  * table read it is still refused by name rather than slipped past.
- *
- * ## Letting go is still written here
- *
- * Removing a contact from a site drops that holder's facet, consent entries,
- * capture attribution and scope tokens — the one change to a facet a holder
- * may make client-direct. Those patches are built here and applied in
- * batches with a per-row fallback, so a refused row is named rather than
- * failing the other 399 with it.
- *
- * Pure apart from the Firestore sentinels the detach carries: the runner takes
- * its writers as ports, so the bar wires `writeBatch`/`updateDoc`/`deleteDoc`
- * in and a spec wires a ledger in.
- */ import { planContactCompanyLink, planContactDetach } from "@aglyn/aglyn";
-import { arrayRemove, deleteField } from "firebase/firestore";
-import { CRM_BULK_WRITE_CHUNK, chunked, runCrmBulkWrites } from "./crm-bulk-writes.js";
-/*
- * The runner and its chunk size are the shared ones (`crm-bulk-writes.ts`,
- * AGL-2621), under the names this module has always exported: the batch
- * cap and the per-row fallback are about Firestore, not about people.
- */ export const CONTACT_BULK_WRITE_CHUNK = CRM_BULK_WRITE_CHUNK;
-export { chunked };
+ */ import { planContactCompanyLink } from "@aglyn/aglyn";
 /** The record page's cap on a holder's tags — the same number, so the two agree. */ export const CONTACT_TAGS_CAP = 20;
 /** The address a report names a row by, falling back to the id. */ export const contactBulkAddressOf = (row)=>String(row.email || row.$id);
 /**
@@ -134,63 +112,5 @@ export { chunked };
         skipped
     };
 }
-/**
- * Let every row go — the record page's DELETE IS A DETACH, per row.
- *
- * `planContactDetach` decides for each document whether this holder is the
- * last one (delete) or one of several (drop this group's facet, consent
- * entries, capture attribution and scope tokens). The `arrayRemove` and
- * `deleteField` sentinels are the ones the record page writes, so a row
- * detached here and a row detached there are the same document afterwards.
- *
- * ⛔ Not the erasure path: a privacy erasure removes the person everywhere
- * regardless of holders.
- */ export function planDetach(rows, group, nowMs) {
-    return {
-        writes: rows.map((row)=>{
-            var _row_visibleTo;
-            // Only the holder tokens reach the plan: that is all it reads, and a
-            // projected table row carries fields the detach must not consult.
-            const plan = planContactDetach({
-                visibleTo: (_row_visibleTo = row.visibleTo) != null ? _row_visibleTo : []
-            }, group);
-            if (plan.action === 'delete') {
-                return {
-                    id: row.$id,
-                    email: contactBulkAddressOf(row),
-                    kind: 'delete'
-                };
-            }
-            return {
-                id: row.$id,
-                email: contactBulkAddressOf(row),
-                kind: 'update',
-                data: _extends({}, Object.fromEntries(plan.remove.map((path)=>[
-                        path,
-                        deleteField()
-                    ])), {
-                    visibleTo: arrayRemove(...plan.removeTokens),
-                    capturedByHostIds: arrayRemove(...plan.removeHostIds),
-                    updatedAt: new Date(nowMs)
-                })
-            };
-        }),
-        skipped: []
-    };
-}
-/**
- * The shared runner over contact writes — batched, with a per-row pass for
- * the chunk that failed (`crm-bulk-writes.ts`) — reporting by ADDRESS, which
- * is the name a contacts report lists a row under.
- */ export async function runContactBulkWrites(writers, writes, chunkSize = CONTACT_BULK_WRITE_CHUNK) {
-    const outcome = await runCrmBulkWrites(writers, writes, (write)=>write.email, chunkSize);
-    return {
-        done: outcome.done,
-        refused: outcome.refused.map((row)=>({
-                email: row.label,
-                error: row.error
-            }))
-    };
-}
 
 //# sourceMappingURL=contacts-bulk-writes.js.map

package/src/lib/model/contacts-bulk-writes.js.map

@@ -1 +1 @@
-{"version":3,"sources":["../../../../../../../libs/plugins/crm/src/lib/model/contacts-bulk-writes.ts"],"sourcesContent":["/**\n * @license\n * Copyright 2026 Aglyn LLC\n *\n * Licensed under the Apache License, Version 2.0 (the \"License\");\n * you may not use this file except in compliance with the License.\n * You may obtain a copy of the License at\n *\n *   http://www.apache.org/licenses/LICENSE-2.0\n *\n * Unless required by applicable law or agreed to in writing, software\n * distributed under the License is distributed on an \"AS IS\" BASIS,\n * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n * See the License for the specific language governing permissions and\n * limitations under the License.\n */\n\n/**\n * What a bulk action on the contacts table does to each row (AGL-2603, and\n * through the server since AGL-2804).\n *\n * The bar over the table offers one act over many rows — tag them, hand them\n * to an owner, file them under a company, move them along the funnel, let\n * them go. A contact's tags, owner and company are one holder's records in\n * that holder's facet, and a facet is the server's to write: the rules cannot\n * tell one field of it from another, so they leave a client nothing there\n * but letting a holder go. So every act but that one is a request to a route,\n * and what this module decides is WHICH rows the request names — the bar\n * holds only the controls.\n *\n * ## A row with nothing to change is left out, and a row that cannot be\n * ## reached is REPORTED\n *\n * A row already carrying the tag, or already at the company, is left out of\n * the request silently: the route would write nothing, and a report saying\n * so would be noise. A row already at the tag cap, or whose company link the\n * table could not project, is left out and named by address — a count of\n * \"398 of 400\" is a number nobody can act on; an address is. The route asks\n * the same questions of the stored document, so a row that changed since the\n * table read it is still refused by name rather than slipped past.\n *\n * ## Letting go is still written here\n *\n * Removing a contact from a site drops that holder's facet, consent entries,\n * capture attribution and scope tokens — the one change to a facet a holder\n * may make client-direct. Those patches are built here and applied in\n * batches with a per-row fallback, so a refused row is named rather than\n * failing the other 399 with it.\n *\n * Pure apart from the Firestore sentinels the detach carries: the runner takes\n * its writers as ports, so the bar wires `writeBatch`/`updateDoc`/`deleteDoc`\n * in and a spec wires a ledger in.\n */\n\nimport {\n  type ContactCompanyLinkState,\n  planContactCompanyLink,\n  planContactDetach,\n} from '@aglyn/aglyn'\nimport { arrayRemove, deleteField } from 'firebase/firestore'\nimport {\n  CRM_BULK_WRITE_CHUNK,\n  chunked,\n  runCrmBulkWrites,\n} from './crm-bulk-writes'\n\n/*\n * The runner and its chunk size are the shared ones (`crm-bulk-writes.ts`,\n * AGL-2621), under the names this module has always exported: the batch\n * cap and the per-row fallback are about Firestore, not about people.\n */\nexport const CONTACT_BULK_WRITE_CHUNK = CRM_BULK_WRITE_CHUNK\nexport { chunked }\n\n/** The record page's cap on a holder's tags — the same number, so the two agree. */\nexport const CONTACT_TAGS_CAP = 20\n\n/** As much of a table row as a bulk act reads. */\nexport interface ContactBulkRow {\n  $id: string\n  email?: string\n  /**\n   * The holder this row was flattened through (`ContactRecord.groupId`),\n   * which a cross-holder reader — the organization-level list (AGL-2630) —\n   * writes each row through, since no one viewing group covers every\n   * selected person there.\n   */\n  groupId?: string\n  /** A site of that holder's group — where the row's stage is moved (`ContactRecord.holderHostId`). */\n  holderHostId?: string\n  /** THIS holder's tags, already read through the facet by the table. */\n  tags?: string[]\n  /** The holder tokens — what `planContactDetach` counts. */\n  visibleTo?: string[]\n  /** The link state the company planner reads — see `ContactRecord.companyLink`. */\n  companyLink?: ContactCompanyLinkState\n}\n\n/** One client-direct write to one contact document, named by the address it is about. */\nexport type ContactBulkWrite = { id: string; email: string } & (\n  | { kind: 'update'; data: Record<string, unknown> }\n  | { kind: 'delete' }\n)\n\n/** A row an action deliberately left alone, and why. */\nexport interface ContactBulkSkip {\n  email: string\n  reason: string\n}\n\n/** What a letting-go wants written, and what it declined to. */\nexport interface ContactBulkPlan {\n  writes: ContactBulkWrite[]\n  skipped: ContactBulkSkip[]\n}\n\n/** The rows a request to a route names, and the rows it leaves out by name. */\nexport interface ContactBulkSelection {\n  rows: ContactBulkRow[]\n  skipped: ContactBulkSkip[]\n}\n\n/** The address a report names a row by, falling back to the id. */\nexport const contactBulkAddressOf = (row: ContactBulkRow): string =>\n  String(row.email || row.$id)\n\n/**\n * A typed tag as the record page stores one, or `null` when nothing survives.\n *\n * Lowercased and trimmed because the record page lowercases and trims, and\n * `contactMatchesSegment` compares lowercased — a bulk \"VIP\" beside a typed\n * \"vip\" would otherwise be two tags on one list and one filter finding half\n * of them.\n */\nexport function normalizeBulkTag(input: string): string | null {\n  const tag = input.trim().toLowerCase().slice(0, 60)\n  return tag || null\n}\n\n/**\n * The rows adding one tag reaches.\n *\n * A row that already carries the tag is left out silently. A row already at\n * the cap without it is left out AND reported: the cap is the record page's,\n * and a bulk path that slipped past it would leave a tag the page's next\n * save silently drops.\n */\nexport function planAddTag(\n  rows: readonly ContactBulkRow[],\n  tag: string,\n): ContactBulkSelection {\n  const reached: ContactBulkRow[] = []\n  const skipped: ContactBulkSkip[] = []\n  for (const row of rows) {\n    const held = (row.tags ?? []).map((held) => held.toLowerCase())\n    if (held.includes(tag)) continue\n    if (held.length >= CONTACT_TAGS_CAP) {\n      skipped.push({\n        email: contactBulkAddressOf(row),\n        reason: `already has ${CONTACT_TAGS_CAP} tags`,\n      })\n      continue\n    }\n    reached.push(row)\n  }\n  return { rows: reached, skipped }\n}\n\n/** The rows taking one tag off reaches: the ones that carry it. */\nexport function planRemoveTag(\n  rows: readonly ContactBulkRow[],\n  tag: string,\n): ContactBulkSelection {\n  return {\n    rows: rows.filter((row) =>\n      (row.tags ?? []).map((held) => held.toLowerCase()).includes(tag),\n    ),\n    skipped: [],\n  }\n}\n\n/**\n * The rows filing under one company reaches — or unlinking, with `null`\n * (AGL-2613).\n *\n * A row already where it was asked to be is left out silently, as a row\n * already tagged is. A row the table could not project a link state for is\n * left out AND named: whether its link would change cannot be told, and the\n * request is not the place to guess.\n */\nexport function planSetCompany(\n  rows: readonly ContactBulkRow[],\n  companyId: string | null,\n): ContactBulkSelection {\n  const reached: ContactBulkRow[] = []\n  const skipped: ContactBulkSkip[] = []\n  for (const row of rows) {\n    if (!row.companyLink) {\n      skipped.push({\n        email: contactBulkAddressOf(row),\n        reason: 'its company link could not be read',\n      })\n      continue\n    }\n    if (planContactCompanyLink(row.companyLink, companyId)) reached.push(row)\n  }\n  return { rows: reached, skipped }\n}\n\n/**\n * Let every row go — the record page's DELETE IS A DETACH, per row.\n *\n * `planContactDetach` decides for each document whether this holder is the\n * last one (delete) or one of several (drop this group's facet, consent\n * entries, capture attribution and scope tokens). The `arrayRemove` and\n * `deleteField` sentinels are the ones the record page writes, so a row\n * detached here and a row detached there are the same document afterwards.\n *\n * ⛔ Not the erasure path: a privacy erasure removes the person everywhere\n * regardless of holders.\n */\nexport function planDetach(\n  rows: readonly ContactBulkRow[],\n  group: { groupId: string; hostIds: readonly string[] },\n  nowMs: number,\n): ContactBulkPlan {\n  return {\n    writes: rows.map((row) => {\n      // Only the holder tokens reach the plan: that is all it reads, and a\n      // projected table row carries fields the detach must not consult.\n      const plan = planContactDetach({ visibleTo: row.visibleTo ?? [] }, group)\n      if (plan.action === 'delete') {\n        return { id: row.$id, email: contactBulkAddressOf(row), kind: 'delete' as const }\n      }\n      return {\n        id: row.$id,\n        email: contactBulkAddressOf(row),\n        kind: 'update' as const,\n        data: {\n          ...Object.fromEntries(plan.remove.map((path) => [path, deleteField()])),\n          visibleTo: arrayRemove(...plan.removeTokens),\n          capturedByHostIds: arrayRemove(...plan.removeHostIds),\n          updatedAt: new Date(nowMs),\n        },\n      }\n    }),\n    skipped: [],\n  }\n}\n\n/** The writers the runner needs — Firestore's, or a spec's ledger. */\nexport interface ContactBulkWriters {\n  /** Apply every write atomically, or throw. */\n  commitBatch: (writes: readonly ContactBulkWrite[]) => Promise<void>\n  /** Apply one write, or throw. */\n  commitOne: (write: ContactBulkWrite) => Promise<void>\n}\n\nexport interface ContactBulkOutcome {\n  /** Rows written. */\n  done: number\n  /** Rows the store refused, by address, with the reason it gave. */\n  refused: Array<{ email: string; error: string }>\n}\n\n/**\n * The shared runner over contact writes — batched, with a per-row pass for\n * the chunk that failed (`crm-bulk-writes.ts`) — reporting by ADDRESS, which\n * is the name a contacts report lists a row under.\n */\nexport async function runContactBulkWrites(\n  writers: ContactBulkWriters,\n  writes: readonly ContactBulkWrite[],\n  chunkSize: number = CONTACT_BULK_WRITE_CHUNK,\n): Promise<ContactBulkOutcome> {\n  const outcome = await runCrmBulkWrites(\n    writers,\n    writes,\n    (write) => write.email,\n    chunkSize,\n  )\n  return {\n    done: outcome.done,\n    refused: outcome.refused.map((row) => ({ email: row.label, error: row.error })),\n  }\n}\n"],"names":["planContactCompanyLink","planContactDetach","arrayRemove","deleteField","CRM_BULK_WRITE_CHUNK","chunked","runCrmBulkWrites","CONTACT_BULK_WRITE_CHUNK","CONTACT_TAGS_CAP","contactBulkAddressOf","row","String","email","$id","normalizeBulkTag","input","tag","trim","toLowerCase","slice","planAddTag","rows","reached","skipped","held","tags","map","includes","length","push","reason","planRemoveTag","filter","planSetCompany","companyId","companyLink","planDetach","group","nowMs","writes","plan","visibleTo","action","id","kind","data","Object","fromEntries","remove","path","removeTokens","capturedByHostIds","removeHostIds","updatedAt","Date","runContactBulkWrites","writers","chunkSize","outcome","write","done","refused","label","error"],"mappings":";AAAA;;;;;;;;;;;;;;;CAeC,GAED;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CAmCC,GAED,SAEEA,sBAAsB,EACtBC,iBAAiB,QACZ,eAAc;AACrB,SAASC,WAAW,EAAEC,WAAW,QAAQ,qBAAoB;AAC7D,SACEC,oBAAoB,EACpBC,OAAO,EACPC,gBAAgB,QACX,uBAAmB;AAE1B;;;;CAIC,GACD,OAAO,MAAMC,2BAA2BH,qBAAoB;AAC5D,SAASC,OAAO,GAAE;AAElB,kFAAkF,GAClF,OAAO,MAAMG,mBAAmB,GAAE;AA+ClC,iEAAiE,GACjE,OAAO,MAAMC,uBAAuB,CAACC,MACnCC,OAAOD,IAAIE,KAAK,IAAIF,IAAIG,GAAG,EAAC;AAE9B;;;;;;;CAOC,GACD,OAAO,SAASC,iBAAiBC,KAAa;IAC5C,MAAMC,MAAMD,MAAME,IAAI,GAAGC,WAAW,GAAGC,KAAK,CAAC,GAAG;IAChD,OAAOH,OAAO;AAChB;AAEA;;;;;;;CAOC,GACD,OAAO,SAASI,WACdC,IAA+B,EAC/BL,GAAW;IAEX,MAAMM,UAA4B,EAAE;IACpC,MAAMC,UAA6B,EAAE;IACrC,KAAK,MAAMb,OAAOW,KAAM;YACRX;QAAd,MAAMc,OAAO,EAACd,YAAAA,IAAIe,IAAI,YAARf,YAAY,EAAE,EAAEgB,GAAG,CAAC,CAACF,OAASA,KAAKN,WAAW;QAC5D,IAAIM,KAAKG,QAAQ,CAACX,MAAM;QACxB,IAAIQ,KAAKI,MAAM,IAAIpB,kBAAkB;YACnCe,QAAQM,IAAI,CAAC;gBACXjB,OAAOH,qBAAqBC;gBAC5BoB,QAAQ,CAAC,YAAY,EAAEtB,iBAAiB,KAAK,CAAC;YAChD;YACA;QACF;QACAc,QAAQO,IAAI,CAACnB;IACf;IACA,OAAO;QAAEW,MAAMC;QAASC;IAAQ;AAClC;AAEA,iEAAiE,GACjE,OAAO,SAASQ,cACdV,IAA+B,EAC/BL,GAAW;IAEX,OAAO;QACLK,MAAMA,KAAKW,MAAM,CAAC,CAACtB;gBAChBA;mBAAD,EAACA,YAAAA,IAAIe,IAAI,YAARf,YAAY,EAAE,EAAEgB,GAAG,CAAC,CAACF,OAASA,KAAKN,WAAW,IAAIS,QAAQ,CAACX;;QAE9DO,SAAS,EAAE;IACb;AACF;AAEA;;;;;;;;CAQC,GACD,OAAO,SAASU,eACdZ,IAA+B,EAC/Ba,SAAwB;IAExB,MAAMZ,UAA4B,EAAE;IACpC,MAAMC,UAA6B,EAAE;IACrC,KAAK,MAAMb,OAAOW,KAAM;QACtB,IAAI,CAACX,IAAIyB,WAAW,EAAE;YACpBZ,QAAQM,IAAI,CAAC;gBACXjB,OAAOH,qBAAqBC;gBAC5BoB,QAAQ;YACV;YACA;QACF;QACA,IAAI9B,uBAAuBU,IAAIyB,WAAW,EAAED,YAAYZ,QAAQO,IAAI,CAACnB;IACvE;IACA,OAAO;QAAEW,MAAMC;QAASC;IAAQ;AAClC;AAEA;;;;;;;;;;;CAWC,GACD,OAAO,SAASa,WACdf,IAA+B,EAC/BgB,KAAsD,EACtDC,KAAa;IAEb,OAAO;QACLC,QAAQlB,KAAKK,GAAG,CAAC,CAAChB;gBAG4BA;YAF5C,qEAAqE;YACrE,kEAAkE;YAClE,MAAM8B,OAAOvC,kBAAkB;gBAAEwC,SAAS,GAAE/B,iBAAAA,IAAI+B,SAAS,YAAb/B,iBAAiB,EAAE;YAAC,GAAG2B;YACnE,IAAIG,KAAKE,MAAM,KAAK,UAAU;gBAC5B,OAAO;oBAAEC,IAAIjC,IAAIG,GAAG;oBAAED,OAAOH,qBAAqBC;oBAAMkC,MAAM;gBAAkB;YAClF;YACA,OAAO;gBACLD,IAAIjC,IAAIG,GAAG;gBACXD,OAAOH,qBAAqBC;gBAC5BkC,MAAM;gBACNC,MAAM,aACDC,OAAOC,WAAW,CAACP,KAAKQ,MAAM,CAACtB,GAAG,CAAC,CAACuB,OAAS;wBAACA;wBAAM9C;qBAAc;oBACrEsC,WAAWvC,eAAesC,KAAKU,YAAY;oBAC3CC,mBAAmBjD,eAAesC,KAAKY,aAAa;oBACpDC,WAAW,IAAIC,KAAKhB;;YAExB;QACF;QACAf,SAAS,EAAE;IACb;AACF;AAiBA;;;;CAIC,GACD,OAAO,eAAegC,qBACpBC,OAA2B,EAC3BjB,MAAmC,EACnCkB,YAAoBlD,wBAAwB;IAE5C,MAAMmD,UAAU,MAAMpD,iBACpBkD,SACAjB,QACA,CAACoB,QAAUA,MAAM/C,KAAK,EACtB6C;IAEF,OAAO;QACLG,MAAMF,QAAQE,IAAI;QAClBC,SAASH,QAAQG,OAAO,CAACnC,GAAG,CAAC,CAAChB,MAAS,CAAA;gBAAEE,OAAOF,IAAIoD,KAAK;gBAAEC,OAAOrD,IAAIqD,KAAK;YAAC,CAAA;IAC9E;AACF"}
+{"version":3,"sources":["../../../../../../../libs/plugins/crm/src/lib/model/contacts-bulk-writes.ts"],"sourcesContent":["/**\n * @license\n * Copyright 2026 Aglyn LLC\n *\n * Licensed under the Apache License, Version 2.0 (the \"License\");\n * you may not use this file except in compliance with the License.\n * You may obtain a copy of the License at\n *\n *   http://www.apache.org/licenses/LICENSE-2.0\n *\n * Unless required by applicable law or agreed to in writing, software\n * distributed under the License is distributed on an \"AS IS\" BASIS,\n * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n * See the License for the specific language governing permissions and\n * limitations under the License.\n */\n\n/**\n * What a bulk action on the contacts table does to each row (AGL-2603, and\n * through the server since AGL-2804).\n *\n * The bar over the table offers one act over many rows — tag them, hand them\n * to an owner, file them under a company, move them along the funnel, let\n * them go. A contact's tags, owner and company are one holder's records in\n * that holder's facet, and a facet is the server's to write: the rules cannot\n * tell one field of it from another. So every act is a request to a route,\n * and what this module decides is WHICH rows the request names — the bar\n * holds only the controls.\n *\n * ## A row with nothing to change is left out, and a row that cannot be\n * ## reached is REPORTED\n *\n * A row already carrying the tag, or already at the company, is left out of\n * the request silently: the route would write nothing, and a report saying\n * so would be noise. A row already at the tag cap, or whose company link the\n * table could not project, is left out and named by address — a count of\n * \"398 of 400\" is a number nobody can act on; an address is. The route asks\n * the same questions of the stored document, so a row that changed since the\n * table read it is still refused by name rather than slipped past.\n */\n\nimport {\n  type ContactCompanyLinkState,\n  planContactCompanyLink,\n} from '@aglyn/aglyn'\n\n/** The record page's cap on a holder's tags — the same number, so the two agree. */\nexport const CONTACT_TAGS_CAP = 20\n\n/** As much of a table row as a bulk act reads. */\nexport interface ContactBulkRow {\n  $id: string\n  email?: string\n  /**\n   * The holder this row was flattened through (`ContactRecord.groupId`),\n   * which a cross-holder reader — the organization-level list (AGL-2630) —\n   * writes each row through, since no one viewing group covers every\n   * selected person there.\n   */\n  groupId?: string\n  /** A site of that holder's group — where the row's stage is moved (`ContactRecord.holderHostId`). */\n  holderHostId?: string\n  /** THIS holder's tags, already read through the facet by the table. */\n  tags?: string[]\n  /** The holder tokens the row is shared with. */\n  visibleTo?: string[]\n  /** The link state the company planner reads — see `ContactRecord.companyLink`. */\n  companyLink?: ContactCompanyLinkState\n}\n\n/** A row an action deliberately left alone, and why. */\nexport interface ContactBulkSkip {\n  email: string\n  reason: string\n}\n\n/** The rows a request to a route names, and the rows it leaves out by name. */\nexport interface ContactBulkSelection {\n  rows: ContactBulkRow[]\n  skipped: ContactBulkSkip[]\n}\n\n/** The address a report names a row by, falling back to the id. */\nexport const contactBulkAddressOf = (row: ContactBulkRow): string =>\n  String(row.email || row.$id)\n\n/**\n * A typed tag as the record page stores one, or `null` when nothing survives.\n *\n * Lowercased and trimmed because the record page lowercases and trims, and\n * `contactMatchesSegment` compares lowercased — a bulk \"VIP\" beside a typed\n * \"vip\" would otherwise be two tags on one list and one filter finding half\n * of them.\n */\nexport function normalizeBulkTag(input: string): string | null {\n  const tag = input.trim().toLowerCase().slice(0, 60)\n  return tag || null\n}\n\n/**\n * The rows adding one tag reaches.\n *\n * A row that already carries the tag is left out silently. A row already at\n * the cap without it is left out AND reported: the cap is the record page's,\n * and a bulk path that slipped past it would leave a tag the page's next\n * save silently drops.\n */\nexport function planAddTag(\n  rows: readonly ContactBulkRow[],\n  tag: string,\n): ContactBulkSelection {\n  const reached: ContactBulkRow[] = []\n  const skipped: ContactBulkSkip[] = []\n  for (const row of rows) {\n    const held = (row.tags ?? []).map((held) => held.toLowerCase())\n    if (held.includes(tag)) continue\n    if (held.length >= CONTACT_TAGS_CAP) {\n      skipped.push({\n        email: contactBulkAddressOf(row),\n        reason: `already has ${CONTACT_TAGS_CAP} tags`,\n      })\n      continue\n    }\n    reached.push(row)\n  }\n  return { rows: reached, skipped }\n}\n\n/** The rows taking one tag off reaches: the ones that carry it. */\nexport function planRemoveTag(\n  rows: readonly ContactBulkRow[],\n  tag: string,\n): ContactBulkSelection {\n  return {\n    rows: rows.filter((row) =>\n      (row.tags ?? []).map((held) => held.toLowerCase()).includes(tag),\n    ),\n    skipped: [],\n  }\n}\n\n/**\n * The rows filing under one company reaches — or unlinking, with `null`\n * (AGL-2613).\n *\n * A row already where it was asked to be is left out silently, as a row\n * already tagged is. A row the table could not project a link state for is\n * left out AND named: whether its link would change cannot be told, and the\n * request is not the place to guess.\n */\nexport function planSetCompany(\n  rows: readonly ContactBulkRow[],\n  companyId: string | null,\n): ContactBulkSelection {\n  const reached: ContactBulkRow[] = []\n  const skipped: ContactBulkSkip[] = []\n  for (const row of rows) {\n    if (!row.companyLink) {\n      skipped.push({\n        email: contactBulkAddressOf(row),\n        reason: 'its company link could not be read',\n      })\n      continue\n    }\n    if (planContactCompanyLink(row.companyLink, companyId)) reached.push(row)\n  }\n  return { rows: reached, skipped }\n}\n"],"names":["planContactCompanyLink","CONTACT_TAGS_CAP","contactBulkAddressOf","row","String","email","$id","normalizeBulkTag","input","tag","trim","toLowerCase","slice","planAddTag","rows","reached","skipped","held","tags","map","includes","length","push","reason","planRemoveTag","filter","planSetCompany","companyId","companyLink"],"mappings":"AAAA;;;;;;;;;;;;;;;CAeC,GAED;;;;;;;;;;;;;;;;;;;;;;CAsBC,GAED,SAEEA,sBAAsB,QACjB,eAAc;AAErB,kFAAkF,GAClF,OAAO,MAAMC,mBAAmB,GAAE;AAmClC,iEAAiE,GACjE,OAAO,MAAMC,uBAAuB,CAACC,MACnCC,OAAOD,IAAIE,KAAK,IAAIF,IAAIG,GAAG,EAAC;AAE9B;;;;;;;CAOC,GACD,OAAO,SAASC,iBAAiBC,KAAa;IAC5C,MAAMC,MAAMD,MAAME,IAAI,GAAGC,WAAW,GAAGC,KAAK,CAAC,GAAG;IAChD,OAAOH,OAAO;AAChB;AAEA;;;;;;;CAOC,GACD,OAAO,SAASI,WACdC,IAA+B,EAC/BL,GAAW;IAEX,MAAMM,UAA4B,EAAE;IACpC,MAAMC,UAA6B,EAAE;IACrC,KAAK,MAAMb,OAAOW,KAAM;YACRX;QAAd,MAAMc,OAAO,EAACd,YAAAA,IAAIe,IAAI,YAARf,YAAY,EAAE,EAAEgB,GAAG,CAAC,CAACF,OAASA,KAAKN,WAAW;QAC5D,IAAIM,KAAKG,QAAQ,CAACX,MAAM;QACxB,IAAIQ,KAAKI,MAAM,IAAIpB,kBAAkB;YACnCe,QAAQM,IAAI,CAAC;gBACXjB,OAAOH,qBAAqBC;gBAC5BoB,QAAQ,CAAC,YAAY,EAAEtB,iBAAiB,KAAK,CAAC;YAChD;YACA;QACF;QACAc,QAAQO,IAAI,CAACnB;IACf;IACA,OAAO;QAAEW,MAAMC;QAASC;IAAQ;AAClC;AAEA,iEAAiE,GACjE,OAAO,SAASQ,cACdV,IAA+B,EAC/BL,GAAW;IAEX,OAAO;QACLK,MAAMA,KAAKW,MAAM,CAAC,CAACtB;gBAChBA;mBAAD,EAACA,YAAAA,IAAIe,IAAI,YAARf,YAAY,EAAE,EAAEgB,GAAG,CAAC,CAACF,OAASA,KAAKN,WAAW,IAAIS,QAAQ,CAACX;;QAE9DO,SAAS,EAAE;IACb;AACF;AAEA;;;;;;;;CAQC,GACD,OAAO,SAASU,eACdZ,IAA+B,EAC/Ba,SAAwB;IAExB,MAAMZ,UAA4B,EAAE;IACpC,MAAMC,UAA6B,EAAE;IACrC,KAAK,MAAMb,OAAOW,KAAM;QACtB,IAAI,CAACX,IAAIyB,WAAW,EAAE;YACpBZ,QAAQM,IAAI,CAAC;gBACXjB,OAAOH,qBAAqBC;gBAC5BoB,QAAQ;YACV;YACA;QACF;QACA,IAAIvB,uBAAuBG,IAAIyB,WAAW,EAAED,YAAYZ,QAAQO,IAAI,CAACnB;IACvE;IACA,OAAO;QAAEW,MAAMC;QAASC;IAAQ;AAClC"}

package/src/lib/model/contacts-csv.d.ts

@@ -37,8 +37,8 @@
  * written, because they are what a merchant reads the file for; they map to
  * "Do not import" on the way back.
  */
-export { CONTACT_CSV_COLUMNS, contactCsvHeader, contactsCsv, type ContactCsvOptions, type ContactCsvRow, } from '@aglyn/aglyn';
-import { type ContactCsvOptions } from '@aglyn/aglyn';
+export { CONTACT_CSV_COLUMNS, contactCsvHeader, contactsCsv, type ContactCsvOptions, type ContactCsvRow, } from './crm-csv';
+import { type ContactCsvOptions } from './crm-csv';
 /**
  * The file the Import drawer hands out to start from: the export's header
  * and nothing under it, so a sheet filled in against it maps itself.

package/src/lib/model/contacts-csv.js

@@ -35,8 +35,8 @@
  * the last interaction and the last campaign engagement — are still
  * written, because they are what a merchant reads the file for; they map to
  * "Do not import" on the way back.
- */ /* The writer moved to `@aglyn/aglyn` under AGL-2662 — see `deals-csv.ts`. */ export { CONTACT_CSV_COLUMNS, contactCsvHeader, contactsCsv } from "@aglyn/aglyn";
-import { contactsCsv } from "@aglyn/aglyn";
+ */ /* The writer is `crm-csv.ts`, shared with the server's export — see `deals-csv.ts`. */ export { CONTACT_CSV_COLUMNS, contactCsvHeader, contactsCsv } from "./crm-csv.js";
+import { contactsCsv } from "./crm-csv.js";
 /**
  * The file the Import drawer hands out to start from: the export's header
  * and nothing under it, so a sheet filled in against it maps itself.

package/src/lib/model/contacts-csv.js.map

@@ -1 +1 @@
-{"version":3,"sources":["../../../../../../../libs/plugins/crm/src/lib/model/contacts-csv.ts"],"sourcesContent":["/**\n * @license\n * Copyright 2026 Aglyn LLC\n *\n * Licensed under the Apache License, Version 2.0 (the \"License\");\n * you may not use this file except in compliance with the License.\n * You may obtain a copy of the License at\n *\n *   http://www.apache.org/licenses/LICENSE-2.0\n *\n * Unless required by applicable law or agreed to in writing, software\n * distributed under the License is distributed on an \"AS IS\" BASIS,\n * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n * See the License for the specific language governing permissions and\n * limitations under the License.\n */\n\n/**\n * The contacts CSV — every CRM column, the same file from the table's\n * Export button and the bulk bar's (AGL-2603, widened by AGL-2621).\n *\n * The selection's export and the table's export are the same file over a\n * different set of rows: the same columns in the same order, the same `|`\n * between multi-valued cells, the same quoting. A second column list would\n * be a second file format for one feature, and the person opening both in\n * a spreadsheet would be the one to find the difference.\n *\n * ## The header row is the import's own vocabulary\n *\n * Each column is headed by a name `guessContactImportMapping` recognizes —\n * \"Job title\", \"Address line 1\", \"Lifecycle stage\", a custom field's label\n * — so an export re-imports without a hand mapping, and the template the\n * Import drawer offers IS this header over no rows. The owner column\n * carries the member's ADDRESS rather than a uid for the same reason: the\n * import resolves an owner by email, and a uid is nothing a spreadsheet\n * can read anyway. The three columns the import has no field for — sources,\n * the last interaction and the last campaign engagement — are still\n * written, because they are what a merchant reads the file for; they map to\n * \"Do not import\" on the way back.\n */\n\n/* The writer moved to `@aglyn/aglyn` under AGL-2662 — see `deals-csv.ts`. */\nexport {\n  CONTACT_CSV_COLUMNS,\n  contactCsvHeader,\n  contactsCsv,\n  type ContactCsvOptions,\n  type ContactCsvRow,\n} from '@aglyn/aglyn'\n\nimport { contactsCsv, type ContactCsvOptions } from '@aglyn/aglyn'\n\n/**\n * The file the Import drawer hands out to start from: the export's header\n * and nothing under it, so a sheet filled in against it maps itself.\n */\nexport function contactImportTemplateCsv(\n  customFields: ContactCsvOptions['customFields'] = [],\n): string {\n  return contactsCsv([], { customFields })\n}\n\n/**\n * Hand the browser a file to save.\n *\n * The one writer that stayed: it touches `URL.createObjectURL` and\n * `document`, so it belongs to the browser and not to the library a server\n * route imports.\n */\nexport function downloadTextFile(name: string, mime: string, text: string): void {\n  const url = URL.createObjectURL(new Blob([text], { type: mime }))\n  const anchor = document.createElement('a')\n  anchor.href = url\n  anchor.download = name\n  anchor.click()\n  URL.revokeObjectURL(url)\n}\n"],"names":["CONTACT_CSV_COLUMNS","contactCsvHeader","contactsCsv","contactImportTemplateCsv","customFields","downloadTextFile","name","mime","text","url","URL","createObjectURL","Blob","type","anchor","document","createElement","href","download","click","revokeObjectURL"],"mappings":"AAAA;;;;;;;;;;;;;;;CAeC,GAED;;;;;;;;;;;;;;;;;;;;;;CAsBC,GAED,2EAA2E,GAC3E,SACEA,mBAAmB,EACnBC,gBAAgB,EAChBC,WAAW,QAGN,eAAc;AAErB,SAASA,WAAW,QAAgC,eAAc;AAElE;;;CAGC,GACD,OAAO,SAASC,yBACdC,eAAkD,EAAE;IAEpD,OAAOF,YAAY,EAAE,EAAE;QAAEE;IAAa;AACxC;AAEA;;;;;;CAMC,GACD,OAAO,SAASC,iBAAiBC,IAAY,EAAEC,IAAY,EAAEC,IAAY;IACvE,MAAMC,MAAMC,IAAIC,eAAe,CAAC,IAAIC,KAAK;QAACJ;KAAK,EAAE;QAAEK,MAAMN;IAAK;IAC9D,MAAMO,SAASC,SAASC,aAAa,CAAC;IACtCF,OAAOG,IAAI,GAAGR;IACdK,OAAOI,QAAQ,GAAGZ;IAClBQ,OAAOK,KAAK;IACZT,IAAIU,eAAe,CAACX;AACtB"}
+{"version":3,"sources":["../../../../../../../libs/plugins/crm/src/lib/model/contacts-csv.ts"],"sourcesContent":["/**\n * @license\n * Copyright 2026 Aglyn LLC\n *\n * Licensed under the Apache License, Version 2.0 (the \"License\");\n * you may not use this file except in compliance with the License.\n * You may obtain a copy of the License at\n *\n *   http://www.apache.org/licenses/LICENSE-2.0\n *\n * Unless required by applicable law or agreed to in writing, software\n * distributed under the License is distributed on an \"AS IS\" BASIS,\n * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n * See the License for the specific language governing permissions and\n * limitations under the License.\n */\n\n/**\n * The contacts CSV — every CRM column, the same file from the table's\n * Export button and the bulk bar's (AGL-2603, widened by AGL-2621).\n *\n * The selection's export and the table's export are the same file over a\n * different set of rows: the same columns in the same order, the same `|`\n * between multi-valued cells, the same quoting. A second column list would\n * be a second file format for one feature, and the person opening both in\n * a spreadsheet would be the one to find the difference.\n *\n * ## The header row is the import's own vocabulary\n *\n * Each column is headed by a name `guessContactImportMapping` recognizes —\n * \"Job title\", \"Address line 1\", \"Lifecycle stage\", a custom field's label\n * — so an export re-imports without a hand mapping, and the template the\n * Import drawer offers IS this header over no rows. The owner column\n * carries the member's ADDRESS rather than a uid for the same reason: the\n * import resolves an owner by email, and a uid is nothing a spreadsheet\n * can read anyway. The three columns the import has no field for — sources,\n * the last interaction and the last campaign engagement — are still\n * written, because they are what a merchant reads the file for; they map to\n * \"Do not import\" on the way back.\n */\n\n/* The writer is `crm-csv.ts`, shared with the server's export — see `deals-csv.ts`. */\nexport {\n  CONTACT_CSV_COLUMNS,\n  contactCsvHeader,\n  contactsCsv,\n  type ContactCsvOptions,\n  type ContactCsvRow,\n} from './crm-csv'\n\nimport { contactsCsv, type ContactCsvOptions } from './crm-csv'\n\n/**\n * The file the Import drawer hands out to start from: the export's header\n * and nothing under it, so a sheet filled in against it maps itself.\n */\nexport function contactImportTemplateCsv(\n  customFields: ContactCsvOptions['customFields'] = [],\n): string {\n  return contactsCsv([], { customFields })\n}\n\n/**\n * Hand the browser a file to save.\n *\n * The one writer that stayed: it touches `URL.createObjectURL` and\n * `document`, so it belongs to the browser and not to the library a server\n * route imports.\n */\nexport function downloadTextFile(name: string, mime: string, text: string): void {\n  const url = URL.createObjectURL(new Blob([text], { type: mime }))\n  const anchor = document.createElement('a')\n  anchor.href = url\n  anchor.download = name\n  anchor.click()\n  URL.revokeObjectURL(url)\n}\n"],"names":["CONTACT_CSV_COLUMNS","contactCsvHeader","contactsCsv","contactImportTemplateCsv","customFields","downloadTextFile","name","mime","text","url","URL","createObjectURL","Blob","type","anchor","document","createElement","href","download","click","revokeObjectURL"],"mappings":"AAAA;;;;;;;;;;;;;;;CAeC,GAED;;;;;;;;;;;;;;;;;;;;;;CAsBC,GAED,qFAAqF,GACrF,SACEA,mBAAmB,EACnBC,gBAAgB,EAChBC,WAAW,QAGN,eAAW;AAElB,SAASA,WAAW,QAAgC,eAAW;AAE/D;;;CAGC,GACD,OAAO,SAASC,yBACdC,eAAkD,EAAE;IAEpD,OAAOF,YAAY,EAAE,EAAE;QAAEE;IAAa;AACxC;AAEA;;;;;;CAMC,GACD,OAAO,SAASC,iBAAiBC,IAAY,EAAEC,IAAY,EAAEC,IAAY;IACvE,MAAMC,MAAMC,IAAIC,eAAe,CAAC,IAAIC,KAAK;QAACJ;KAAK,EAAE;QAAEK,MAAMN;IAAK;IAC9D,MAAMO,SAASC,SAASC,aAAa,CAAC;IACtCF,OAAOG,IAAI,GAAGR;IACdK,OAAOI,QAAQ,GAAGZ;IAClBQ,OAAOK,KAAK;IACZT,IAAIU,eAAe,CAACX;AACtB"}

package/src/lib/model/crm-csv.d.ts

@@ -0,0 +1,239 @@
+/**
+ * @license
+ * Copyright 2026 Aglyn LLC
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *   http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+/**
+ * EVERY CRM CSV, WRITTEN ONCE (AGL-2662).
+ *
+ * ## Why these moved out of the plugin
+ *
+ * Each CRM section already wrote a file from the rows it had LOADED, and
+ * the whole-collection export writes the same file from the server over
+ * every row there is. The console app may not import a feature plugin —
+ * plugins reach an app only through the generated loader manifests — so a
+ * server route could not have called the plugin's writers, and writing a
+ * second set would have given one feature two file formats, discovered by
+ * whichever merchant opened both in a spreadsheet.
+ *
+ * So the columns and the cells live here, in the library both sides may
+ * import. The plugin's `contacts-csv.ts`, `companies-csv.ts`,
+ * `deals-csv.ts`, `tasks-csv.ts` and `leads-csv.ts` keep their names and
+ * delegate: the sections import what they always did, and there is one
+ * implementation under all of it. Their specs are unchanged and are the
+ * proof that the move changed no byte of any file.
+ *
+ * ## What a resolver is for, and what an absent one means
+ *
+ * A stored document holds ids — an owner uid, a pipeline id, a linked
+ * record id — and a spreadsheet cannot read an id. Each writer therefore
+ * takes optional resolvers, and every one of them degrades the same way:
+ * ABSENT resolver writes the id itself rather than an empty cell, because
+ * a row whose owner is a uid is still a row whose owner can be found, and
+ * a blank is a claim that nobody owns it.
+ *
+ * The owner columns carry an ADDRESS where they can, which is also what
+ * the contacts and companies imports resolve an owner by — so an exported
+ * file re-imports without a hand mapping.
+ */
+import type { AglynPostalAddress } from '@aglyn/aglyn';
+import { type ContactCustomValue, type ContactFieldDefinition, type ContactLifecycleStage, type CrmCompany, type CrmLeadFields, type CrmTask } from '@aglyn/aglyn/app-utils/crm';
+/** Cents as a major-unit decimal — `125000` → `1250.00`; `''` for none. */
+export declare function csvAmount(cents: number | null | undefined): string;
+/** An epoch as an ISO timestamp, or `''` for none. */
+export declare function csvInstant(ms: number | null | undefined): string;
+/**
+ * An epoch as a local calendar DATE — `YYYY-MM-DD`.
+ *
+ * For the fields that are a day rather than a moment: an expected close is
+ * stored at local noon precisely so the day survives a timezone round trip,
+ * and writing it as an instant would hand a spreadsheet a time nobody
+ * chose. The same arithmetic the deal drawer's date field uses, so the
+ * file and the field read one day.
+ */
+export declare function csvLocalDate(ms: number | null | undefined): string;
+/** Epoch millis or a Firestore timestamp as an ISO instant, or `''`. */
+export declare function csvStoredInstant(value: unknown): string;
+/** As much of a projected contact row as the file reads. */
+export interface ContactCsvRow {
+    email?: string;
+    name?: string;
+    phone?: string;
+    jobTitle?: string;
+    companyName?: string;
+    ownerUid?: string;
+    lifecycleStage?: ContactLifecycleStage | '';
+    address?: AglynPostalAddress | null;
+    tags?: string[];
+    sources?: Record<string, unknown>;
+    interactions?: Array<{
+        atMs: number;
+    }>;
+    /** The last open or click on one of the site's campaigns (AGL-2616). */
+    lastEmailEngagementAtMs?: number;
+    notes?: string;
+    /** This holder's custom values, keyed by definition key. */
+    custom?: Record<string, ContactCustomValue>;
+}
+export interface ContactCsvOptions {
+    /**
+     * The owner's address for a stored uid — what the import resolves an
+     * owner by. Absent, the uid is written as it is.
+     */
+    ownerEmail?: (uid: string) => string;
+    /** The org's custom fields, one column each, headed by the field's label. */
+    customFields?: readonly Pick<ContactFieldDefinition, 'key' | 'label'>[];
+}
+/**
+ * The standard columns, in order, headed as the import reads them. A
+ * custom field's column follows these, headed by its label.
+ */
+export declare const CONTACT_CSV_COLUMNS: readonly ["Email", "Name", "Phone", "Job title", "Company", "Owner", "Lifecycle stage", "Address line 1", "Address line 2", "City", "State", "Postal code", "Country", "Tags", "Sources", "Last interaction", "Last engaged", "Notes"];
+/** The header row, with one column per custom field after the standard ones. */
+export declare function contactCsvHeader(customFields?: ContactCsvOptions['customFields']): string[];
+/** One contact's cells, in the header's order. */
+export declare function contactCsvCells(contact: ContactCsvRow, options?: ContactCsvOptions): unknown[];
+/** The whole file, header first. */
+export declare function contactsCsv(rows: readonly ContactCsvRow[], options?: ContactCsvOptions): string;
+/** As much of a company row as the file reads. */
+export type CompanyCsvRow = Partial<Pick<CrmCompany, 'name' | 'domain' | 'website' | 'phone' | 'industry' | 'ownerUid' | 'tags' | 'notes' | 'contactsCount'>> & {
+    address?: AglynPostalAddress | null;
+};
+export interface CompanyCsvOptions {
+    /** The owner's address for a stored uid; absent, the uid is written. */
+    ownerEmail?: (uid: string) => string;
+}
+/** The columns, in order, headed as the import reads them. */
+export declare const COMPANY_CSV_COLUMNS: readonly ["Company", "Domain", "Website", "Phone", "Industry", "Owner", "Address line 1", "Address line 2", "City", "State", "Postal code", "Country", "Tags", "Notes", "Contacts"];
+/** One company's cells, in the header's order. */
+export declare function companyCsvCells(company: CompanyCsvRow, options?: CompanyCsvOptions): unknown[];
+/** The whole file, header first. */
+export declare function companiesCsv(rows: readonly CompanyCsvRow[], options?: CompanyCsvOptions): string;
+/** The currency a deal with no code of its own is read in. */
+export declare const CSV_DEFAULT_DEAL_CURRENCY = "usd";
+/** How a deal's `status` reads in a file. */
+export declare const CSV_DEAL_STATUS_LABELS: Record<string, string>;
+/** As much of a deal row as the file reads. */
+export interface DealCsvRow {
+    title?: string;
+    pipelineId?: string;
+    stageId?: string;
+    status?: string;
+    amountCents?: number;
+    currency?: string;
+    ownerUid?: string;
+    expectedCloseAtMs?: number | null;
+    closedAtMs?: number | null;
+    contactName?: string;
+    companyName?: string;
+    lostReason?: string;
+    notes?: string;
+}
+export interface DealCsvOptions {
+    /** The pipeline's name for a stored id; absent, the id is written. */
+    pipelineName?: (pipelineId: string) => string | undefined;
+    /** The stage's name within its pipeline; absent, the stage id is written. */
+    stageName?: (pipelineId: string, stageId: string) => string | undefined;
+    /** The owner's address for a stored uid; absent, the uid is written. */
+    ownerEmail?: (uid: string) => string;
+}
+export declare const DEAL_CSV_COLUMNS: readonly ["Title", "Pipeline", "Stage", "Amount", "Currency", "Owner", "Expected close", "Status", "Contact", "Company", "Closed", "Lost reason", "Notes"];
+/** One deal's cells, in the header's order. */
+export declare function dealCsvCells(deal: DealCsvRow, options?: DealCsvOptions): unknown[];
+/** The whole file, header first. */
+export declare function dealsCsv(rows: readonly DealCsvRow[], options?: DealCsvOptions): string;
+/** As much of a task row as the file reads. */
+export type TaskCsvRow = Partial<Pick<CrmTask, 'title' | 'kind' | 'priority' | 'status' | 'dueAtMs' | 'completedAtMs' | 'assigneeUid' | 'contactId' | 'companyId' | 'dealId' | 'notes'>>;
+export interface TaskCsvOptions {
+    /** The assignee's address for a stored uid; absent, the uid is written. */
+    assigneeEmail?: (uid: string) => string;
+    /** What a linked record is called; absent, the id is written. */
+    recordName?: (kind: 'contact' | 'company' | 'deal', id: string) => string | undefined;
+}
+export declare const TASK_CSV_COLUMNS: readonly ["Title", "Kind", "Priority", "Status", "Due", "Assignee", "Contact", "Company", "Deal", "Completed", "Notes"];
+/** One task's cells, in the header's order. */
+export declare function taskCsvCells(task: TaskCsvRow, options?: TaskCsvOptions): unknown[];
+/** The whole file, header first. */
+export declare function tasksCsv(rows: readonly TaskCsvRow[], options?: TaskCsvOptions): string;
+/** How one capture surface reads in a file. */
+export declare function csvLeadSourceLabel(source: string): string;
+/** Every surface that produced a capture — the array, or the older single field. */
+export declare function csvLeadSources(lead: Record<string, unknown>): string[];
+/** As much of a lead row as the file reads. */
+export type LeadCsvRow = Record<string, unknown> & Pick<CrmLeadFields, 'status' | 'ownerUid' | 'notes' | 'unqualifiedReason' | 'convertedAtMs' | 'company' | 'jobTitle' | 'phone' | 'website' | 'address' | 'tags' | 'leadSource'> & {
+    /** The site the lead lives under — what the `Site` column names. */
+    hostId?: string;
+};
+export interface LeadCsvOptions {
+    /** The owner's address for a stored uid; absent, the uid is written. */
+    ownerEmail?: (uid: string) => string;
+    /**
+     * The site's name for a row's `hostId`. Given, the file carries a `Site`
+     * column — the organization-level file; absent, it does not.
+     */
+    siteName?: (hostId: string) => string | undefined;
+}
+/**
+ * The columns every leads file carries, in the list's order.
+ *
+ * The lead's own profile (AGL-3231) — company, title, phone, website, the
+ * address parts, tags and the lead source — sits where the contacts file
+ * keeps the same facts, under the same headers, so one spreadsheet of
+ * people reads into either import with the same mapping.
+ */
+export declare const LEAD_CSV_COLUMNS: readonly ["Email", "Name", "Company", "Job title", "Phone", "Website", "Status", "Owner", "Lead source", "Sources", "First seen", "Last seen", "Captures", "Address line 1", "Address line 2", "City", "State", "Postal code", "Country", "Tags", "Unqualified reason", "Converted", "Notes"];
+/** The header row: the standard columns, with `Site` after `Owner` at the org level. */
+export declare function leadCsvHeader(options?: LeadCsvOptions): string[];
+/** One lead's cells, in the header's order. */
+export declare function leadCsvCells(lead: LeadCsvRow, options?: LeadCsvOptions): unknown[];
+/** The whole file, header first. */
+export declare function leadsCsv(rows: readonly LeadCsvRow[], options?: LeadCsvOptions): string;
+/** The five tables a whole-collection export can be asked for. */
+export declare const CRM_EXPORT_RESOURCES: readonly ["contacts", "companies", "deals", "tasks", "leads"];
+export type CrmExportResource = (typeof CRM_EXPORT_RESOURCES)[number];
+export declare function isCrmExportResource(value: unknown): value is CrmExportResource;
+/** Every resolver any of the five writers takes, in one bag. */
+export type CrmExportOptions = ContactCsvOptions & DealCsvOptions & TaskCsvOptions & LeadCsvOptions & {
+    assigneeEmail?: (uid: string) => string;
+};
+/** The header row one resource's file carries. */
+export declare function crmExportHeader(resource: CrmExportResource, options?: CrmExportOptions): string[];
+/**
+ * One stored document's cells for one resource's file.
+ *
+ * The bridge the server route stands on: it holds a raw document rather
+ * than the projected row a list built, and this is the one place that
+ * turns the first into the second — so the streamed file and the file the
+ * Export button writes are the same columns filled the same way.
+ */
+export declare function crmExportCells(resource: CrmExportResource, document: Record<string, unknown>, options?: CrmExportOptions): unknown[];
+/** What the contacts file needs to know about the holder it is written for. */
+export interface CrmExportContactHolder {
+    /** The consent group whose facet the file reads — a facet is per holder. */
+    groupId: string;
+    /** The sites in that group, which bound the timeline the file reads. */
+    hostIds: readonly string[];
+}
+/**
+ * A stored contact document as the contacts file reads it.
+ *
+ * The projection the Contacts list makes before it writes its own file,
+ * spelled here so the server can make the same one: a contact document
+ * holds one facet per HOLDER, and every field the file writes but the
+ * address itself — the phone, the company, the tags, the notes, the
+ * lifecycle stage, the custom values — lives in the viewing holder's facet.
+ * Reading them off the top of the document is how one business's file ends
+ * up carrying another's notes.
+ */
+export declare function contactCsvRowFromDoc(document: Record<string, unknown>, holder: CrmExportContactHolder): ContactCsvRow;