The Report Model

The report model is the heart of ZiWound — the war crimes documentation entity. It's the deepest relation graph in the app (11 relations) and the model with the most acts. Source: back/models/report.ts.

Pure fields

export const report_status_array = ["Pending", "Approved", "Rejected", "InReview"] as const;
export const report_status_enum = enums(report_status_array);

export const report_pure = {
  title: string(),
  description: string(),
  conflict_date: optional(coerce(date(), string(), (v) => new Date(v))),
  report_status: defaulted(coerce(report_status_enum, string(), (v) => v), "Pending"),
  point: geoJSONStruct("Point"),        // { type: "Point", coordinates: [lng, lat] }
  ...createUpdateAt,
};

Field	Type	Notes
title	string()	text-indexed
description	string()	the incident narrative
conflict_date	date	coerced from ISO string
report_status	enums(["Pending","Approved","Rejected","InReview"])	defaulted to Pending — new submissions start unapproved
point	geoJSONStruct("Point")	GeoJSON point for map display
createdAt / updatedAt	spread from createUpdateAt

Relations (the full graph)

export const report_relations = {
  reporter: {
    schemaName: "user", type: "single" as RelationDataType,
    optional: false, excludes: user_excludes,
    relatedRelations: { reports: { type: "multiple" as RelationDataType, limit: 100 } },
  },
  documents: {
    schemaName: "document", type: "multiple" as RelationDataType,
    excludes: document_excludes,
    relatedRelations: { reports: { type: "multiple" as RelationDataType, limit: 20 } },
  },
  hostileCountries: {
    schemaName: "country", type: "multiple" as RelationDataType,
    excludes: location_excludes,
    relatedRelations: { hostileReports: { type: "multiple" as RelationDataType, limit: 50 } },
  },
  attackedCountries: {
    schemaName: "country", type: "multiple" as RelationDataType,
    excludes: location_excludes,
    relatedRelations: { attackedReports: { type: "multiple" as RelationDataType, limit: 50 } },
  },
  attackedProvinces: {
    schemaName: "province", type: "multiple" as RelationDataType,
    excludes: location_excludes,
    relatedRelations: { attackedReports: { type: "multiple" as RelationDataType, limit: 50 } },
  },
  attackedCities: {
    schemaName: "city", type: "multiple" as RelationDataType,
    excludes: location_excludes,
    relatedRelations: { attackedReports: { type: "multiple" as RelationDataType, limit: 50 } },
  },
  tags: {
    schemaName: "tag", type: "multiple" as RelationDataType,
    excludes: tag_excludes,
    relatedRelations: { reports: { type: "multiple" as RelationDataType, limit: 20 } },
  },
  category: {
    schemaName: "category", type: "single" as RelationDataType,
    excludes: category_excludes,
    relatedRelations: { reports: { type: "multiple" as RelationDataType, limit: 50 } },
  },
  warCriminals: {
    schemaName: "warCriminal", type: "multiple" as RelationDataType,
    excludes: war_criminal_excludes,
    relatedRelations: { reports: { type: "multiple" as RelationDataType, limit: 100 } },
  },
  registrar: {
    schemaName: "user", type: "single" as RelationDataType,
    optional: true, excludes: user_excludes,
    relatedRelations: { registeredReports: { type: "multiple" as RelationDataType, limit: 100 } },
  },
};

Key lessons:

• Different relation kinds to the same schema — reporter (the submitter) and registrar (an admin/editor) are both single → user, but they're separate relations, so a user gets two distinct reverse lists: reports (submitted) and registeredReports (moderated).
• Two country relations — hostileCountries (who attacked) vs attackedCountries (who was attacked). Because relations are named, one report can link the same country in both roles without conflict.
• Per-relation back-references — every multiple relation declares its own limit, so reverse arrays stay bounded: a user's reports caps at 100, a tag's at 20, a category's at 50.

Registration & indexes

export const reports = () =>
  coreApp.odm.newModel("report", report_pure, report_relations, {
    createIndex: { indexSpec: { title: "text", description: "text" }, unique: false },
    excludes: ["password"],
  });

A 2dsphere index on point is created separately so map queries stay fast:

await db.collection("report").createIndex({ point: "2dsphere" });

Acts

The report domain has the largest act set — 17 acts: add, get, gets, update, updateRelations, remove, count, approve, reject, inReview, statistics, getRelated (reports by status/location filters), getRelatedPagination, getRelatedByGeo (geospatial search), getDraft, getDraftCount, getRelatedSitemap.

The status-flow acts (approve/reject/inReview) are a clean example of an act that only flips an enum — see the Authentication & Authorization page for how grantAccess protects them, and Search & Indexes for getRelatedByGeo.

Next: The Document Model.