@topogram/cli 0.3.63 → 0.3.65

package/src/workflows/import-app/db.js

@@ -0,0 +1,538 @@
+// @ts-check
+import path from "node:path";
+
+import { relativeTo } from "../../path-helpers.js";
+import { idHintify, slugify, titleCase } from "../../text-helpers.js";
+import { readJsonIfExists, readTextIfExists } from "../shared.js";
+import { dedupeCandidateRecords, findImportFiles, makeCandidateRecord, selectPreferredImportFiles } from "./shared.js";
+
+/** @param {any} typeName @returns {any} */
+function normalizePrismaType(typeName) {
+  const normalized = String(typeName || "").toLowerCase();
+  switch (normalized) {
+    case "string":
+      return "string";
+    case "int":
+      return "int";
+    case "bigint":
+      return "bigint";
+    case "float":
+      return "float";
+    case "decimal":
+      return "decimal";
+    case "boolean":
+    case "bool":
+      return "boolean";
+    case "datetime":
+      return "datetime";
+    case "bytes":
+      return "bytes";
+    case "json":
+      return "json";
+    default:
+      return typeName;
+  }
+}
+
+/** @param {any} schemaText @returns {any} */
+function parsePrismaSchema(schemaText) {
+  /** @type {any[]} */
+  const enums = [];
+  /** @type {any[]} */
+  const entities = [];
+  /** @type {any[]} */
+  const relations = [];
+  /** @type {any[]} */
+  const indexes = [];
+  const enumNames = new Set();
+  /** @type {any[]} */
+  const modelNames = [];
+
+  for (const match of schemaText.matchAll(/^enum\s+([A-Za-z0-9_]+)\s*\{([\s\S]*?)^\}/gm)) {
+    const [, enumName, body] = match;
+    const values = body
+      .split(/\r?\n/)
+      .map((/** @type {any} */ line) => line.replace(/\/\/.*$/, "").trim())
+      .filter((/** @type {any} */ line) => line && !line.startsWith("@@"))
+      .map((/** @type {any} */ line) => line.split(/\s+/)[0]);
+    enumNames.add(enumName);
+    enums.push({ name: enumName, values });
+  }
+
+  for (const match of schemaText.matchAll(/^model\s+([A-Za-z0-9_]+)\s*\{/gm)) {
+    modelNames.push(match[1]);
+  }
+  const modelNameSet = new Set(modelNames);
+
+  for (const match of schemaText.matchAll(/^model\s+([A-Za-z0-9_]+)\s*\{([\s\S]*?)^\}/gm)) {
+    const [, modelName, body] = match;
+    /** @type {any[]} */
+    const fields = [];
+    /** @type {any[]} */
+    const localIndexes = [];
+    const lines = body
+      .split(/\r?\n/)
+      .map((/** @type {any} */ line) => line.replace(/\/\/.*$/, "").trim())
+      .filter(Boolean);
+
+    for (const line of lines) {
+      if (line.startsWith("@@")) {
+        const indexMatch = line.match(/^@@(unique|index)\(\[([^\]]+)\]/);
+        if (indexMatch) {
+          const [, type, rawFields] = indexMatch;
+          localIndexes.push({
+            id_hint: `index_${slugify(`${modelName}_${rawFields}`)}`,
+            fields: rawFields.split(",").map((/** @type {any} */ field) => field.trim()),
+            unique: type === "unique"
+          });
+        }
+        continue;
+      }
+
+      const fieldMatch = line.match(/^([A-Za-z0-9_]+)\s+([^\s]+)(.*)$/);
+      if (!fieldMatch) {
+        continue;
+      }
+      const [, fieldName, rawTypeToken, remainder] = fieldMatch;
+      const list = rawTypeToken.endsWith("[]");
+      const optional = rawTypeToken.endsWith("?");
+      const baseType = rawTypeToken.replace(/\?|\[\]/g, "");
+      const referencesModel = modelNameSet.has(baseType) && !enumNames.has(baseType);
+      const hasRelationDirective = remainder.includes("@relation(");
+
+      if (referencesModel && hasRelationDirective) {
+        const relationMatch = remainder.match(/@relation\(([^)]*)\)/);
+        const relationArgs = relationMatch?.[1] || "";
+        const fieldsMatch = relationArgs.match(/fields:\s*\[([^\]]+)\]/);
+        const refsMatch = relationArgs.match(/references:\s*\[([^\]]+)\]/);
+        relations.push({
+          from_entity: `entity_${slugify(modelName)}`,
+          to_entity: `entity_${slugify(baseType)}`,
+          relation_field: fieldName,
+          fields: fieldsMatch ? fieldsMatch[1].split(",").map((/** @type {any} */ field) => field.trim()) : [],
+          references: refsMatch ? refsMatch[1].split(",").map((/** @type {any} */ field) => field.trim()) : []
+        });
+        continue;
+      }
+
+      if (referencesModel) {
+        continue;
+      }
+
+      const fieldType = enumNames.has(baseType) ? baseType : normalizePrismaType(baseType);
+      fields.push({
+        name: fieldName,
+        field_type: fieldType,
+        required: !optional && !list,
+        list,
+        unique: /@unique\b/.test(remainder),
+        primary_key: /@id\b/.test(remainder)
+      });
+
+      if (/@unique\b/.test(remainder)) {
+        localIndexes.push({
+          id_hint: `index_${slugify(`${modelName}_${fieldName}_unique`)}`,
+          fields: [fieldName],
+          unique: true
+        });
+      }
+    }
+
+    entities.push({ name: modelName, fields });
+    indexes.push(...localIndexes.map((/** @type {any} */ index) => ({ ...index, entity: `entity_${slugify(modelName)}` })));
+  }
+
+  return { entities, enums, relations, indexes };
+}
+
+/** @param {string} body @returns {any} */
+function splitSqlSegments(body) {
+  return body
+    .split(/,\s*\n/)
+    .map((/** @type {any} */ segment) => segment.trim())
+    .filter(Boolean);
+}
+
+/** @param {any} sqlText @returns {any} */
+function parseSqlSchema(sqlText) {
+  /** @type {any[]} */
+  const entities = [];
+  /** @type {any[]} */
+  const enums = [];
+  /** @type {any[]} */
+  const relations = [];
+  /** @type {any[]} */
+  const indexes = [];
+
+  for (const match of sqlText.matchAll(/CREATE\s+TYPE\s+([A-Za-z0-9_"]+)\s+AS\s+ENUM\s*\(([\s\S]*?)\);/gi)) {
+    const enumName = match[1].replace(/"/g, "");
+    const values = [...match[2].matchAll(/'([^']+)'/g)].map((/** @type {any} */ valueMatch) => valueMatch[1]);
+    enums.push({ name: enumName, values });
+  }
+
+  for (const match of sqlText.matchAll(/CREATE\s+TABLE\s+([A-Za-z0-9_"]+)\s*\(([\s\S]*?)\);/gi)) {
+    const tableName = match[1].replace(/"/g, "");
+    const entityId = `entity_${slugify(tableName.replace(/s$/, ""))}`;
+    /** @type {any[]} */
+    const fields = [];
+    for (const segment of splitSqlSegments(match[2])) {
+      if (/^(PRIMARY\s+KEY|UNIQUE|CONSTRAINT|FOREIGN\s+KEY)/i.test(segment)) {
+        const foreignKeyMatch = segment.match(/FOREIGN\s+KEY\s*\(([^)]+)\)\s+REFERENCES\s+([A-Za-z0-9_"]+)\s*\(([^)]+)\)/i);
+        if (foreignKeyMatch) {
+          relations.push({
+            from_entity: entityId,
+            to_entity: `entity_${slugify(foreignKeyMatch[2].replace(/"/g, "").replace(/s$/, ""))}`,
+            relation_field: foreignKeyMatch[1].replace(/"/g, "").trim(),
+            fields: foreignKeyMatch[1].split(",").map((/** @type {any} */ field) => field.replace(/"/g, "").trim()),
+            references: foreignKeyMatch[3].split(",").map((/** @type {any} */ field) => field.replace(/"/g, "").trim())
+          });
+        }
+        const uniqueMatch = segment.match(/UNIQUE\s*\(([^)]+)\)/i);
+        if (uniqueMatch) {
+          indexes.push({
+            entity: entityId,
+            id_hint: `index_${slugify(`${tableName}_${uniqueMatch[1]}`)}`,
+            fields: uniqueMatch[1].split(",").map((/** @type {any} */ field) => field.replace(/"/g, "").trim()),
+            unique: true
+          });
+        }
+        continue;
+      }
+      const fieldMatch = segment.match(/^"?([A-Za-z0-9_]+)"?\s+([A-Za-z0-9_()[\]]+)(.*)$/i);
+      if (!fieldMatch) {
+        continue;
+      }
+      const [, fieldName, rawType, remainder] = fieldMatch;
+      fields.push({
+        name: fieldName,
+        field_type: normalizePrismaType(rawType.replace(/\(.+$/, "")),
+        required: /NOT\s+NULL/i.test(remainder),
+        list: false,
+        unique: /\bUNIQUE\b/i.test(remainder),
+        primary_key: /\bPRIMARY\s+KEY\b/i.test(remainder)
+      });
+      const inlineReferenceMatch = remainder.match(/REFERENCES\s+([A-Za-z0-9_"]+)\s*\(([^)]+)\)/i);
+      if (inlineReferenceMatch) {
+        relations.push({
+          from_entity: entityId,
+          to_entity: `entity_${slugify(inlineReferenceMatch[1].replace(/"/g, "").replace(/s$/, ""))}`,
+          relation_field: fieldName,
+          fields: [fieldName],
+          references: inlineReferenceMatch[2].split(",").map((/** @type {any} */ field) => field.replace(/"/g, "").trim())
+        });
+      }
+      if (/\bUNIQUE\b/i.test(remainder)) {
+        indexes.push({
+          entity: entityId,
+          id_hint: `index_${slugify(`${tableName}_${fieldName}_unique`)}`,
+          fields: [fieldName],
+          unique: true
+        });
+      }
+    }
+    entities.push({ name: tableName.replace(/s$/, ""), table_name: tableName, fields });
+  }
+
+  for (const match of sqlText.matchAll(/CREATE\s+(UNIQUE\s+)?INDEX\s+([A-Za-z0-9_"]+)\s+ON\s+([A-Za-z0-9_"]+)\s*\(([^)]+)\)/gi)) {
+    indexes.push({
+      entity: `entity_${slugify(match[3].replace(/"/g, "").replace(/s$/, ""))}`,
+      id_hint: `index_${slugify(match[2].replace(/"/g, ""))}`,
+      fields: match[4].split(",").map((/** @type {any} */ field) => field.replace(/"/g, "").trim()),
+      unique: Boolean(match[1])
+    });
+  }
+
+  return { entities, enums, relations, indexes };
+}
+
+/** @param {any} snapshot @returns {any} */
+function parseDbSchemaSnapshot(snapshot) {
+  return {
+    entities: (snapshot.tables || []).map((/** @type {any} */ table) => ({
+      name: table.entity?.name || table.table.replace(/s$/, ""),
+      table_name: table.table,
+      fields: (table.columns || []).map((/** @type {any} */ column) => ({
+        name: column.name,
+        field_type: column.type,
+        required: !column.nullable,
+        list: false,
+        unique: false,
+        primary_key: false
+      }))
+    })),
+    enums: (snapshot.enums || []).map((/** @type {any} */ entry) => ({
+      name: entry.name || entry.id,
+      values: entry.values || []
+    })),
+    relations: (snapshot.tables || []).flatMap((/** @type {any} */ table) =>
+      (table.foreignKeys || []).map((/** @type {any} */ foreignKey) => ({
+        from_entity: table.entity?.id || `entity_${slugify(table.table.replace(/s$/, ""))}`,
+        to_entity: foreignKey.references?.id || foreignKey.reference?.id || `entity_${slugify((foreignKey.references?.table || "").replace(/s$/, ""))}`,
+        relation_field: foreignKey.columns?.[0] || "",
+        fields: foreignKey.columns || [],
+        references: foreignKey.references?.columns || []
+      }))
+    ),
+    indexes: (snapshot.tables || []).flatMap((/** @type {any} */ table) =>
+      (table.indexes || []).map((/** @type {any} */ index) => ({
+        entity: table.entity?.id || `entity_${slugify(table.table.replace(/s$/, ""))}`,
+        id_hint: `index_${slugify(index.name || `${table.table}_${(index.columns || []).join("_")}`)}`,
+        fields: index.columns || [],
+        unique: Boolean(index.unique)
+      }))
+    )
+  };
+}
+
+/** @param {WorkspacePaths} paths @returns {any} */
+export function discoverDbSources(paths) {
+  const allPrismaFiles = findImportFiles(paths, (/** @type {any} */ filePath) => filePath.endsWith(path.join("prisma", "schema.prisma")) || filePath.endsWith("/prisma/schema.prisma"));
+  const allSqlFiles = findImportFiles(paths, (/** @type {any} */ filePath) => filePath.endsWith(".sql") && /(schema|migration|migrations|db)/i.test(filePath));
+  const snapshotFiles = findImportFiles(paths, (/** @type {any} */ filePath) => filePath.endsWith(".db-schema-snapshot.json"));
+  const prismaFiles = selectPreferredImportFiles(paths, allPrismaFiles, "prisma");
+  const schemaSqlFiles = allSqlFiles.filter((/** @type {any} */ filePath) => !/migration/i.test(path.basename(filePath)));
+  const migrationSqlFiles = allSqlFiles.filter((/** @type {any} */ filePath) => /migration/i.test(path.basename(filePath)));
+  const sqlFiles =
+    prismaFiles.length > 0
+      ? []
+      : schemaSqlFiles.length > 0
+        ? selectPreferredImportFiles(paths, schemaSqlFiles, "sql")
+        : selectPreferredImportFiles(paths, migrationSqlFiles, "sql");
+  return { prismaFiles, sqlFiles, snapshotFiles };
+}
+
+/** @param {WorkspacePaths} paths @returns {any} */
+export function collectDbImport(paths) {
+  /** @type {any[]} */
+  const findings = [];
+  /** @type {WorkflowRecord} */
+  const candidates = {
+    entities: [],
+    enums: [],
+    relations: [],
+    indexes: []
+  };
+  const { prismaFiles, sqlFiles, snapshotFiles } = discoverDbSources(paths);
+  let hasPrimarySchemaSource = false;
+
+  for (const filePath of prismaFiles) {
+    const parsed = parsePrismaSchema(readTextIfExists(filePath) || "");
+    const provenance = relativeTo(paths.repoRoot, filePath);
+    hasPrimarySchemaSource = true;
+    findings.push({
+      kind: "prisma_schema",
+      file: provenance,
+      entity_count: parsed.entities.length,
+      enum_count: parsed.enums.length
+    });
+    candidates.entities.push(
+      ...parsed.entities.map((/** @type {any} */ entity) =>
+        makeCandidateRecord({
+          kind: "entity",
+          idHint: `entity_${slugify(entity.name)}`,
+          label: titleCase(entity.name),
+          confidence: "high",
+          sourceKind: "schema",
+          provenance,
+          table_name: slugify(entity.table_name || entity.name),
+          fields: entity.fields
+        })
+      )
+    );
+    candidates.enums.push(
+      ...parsed.enums.map((/** @type {any} */ entry) =>
+        makeCandidateRecord({
+          kind: "enum",
+          idHint: idHintify(entry.name),
+          label: titleCase(entry.name),
+          confidence: "high",
+          sourceKind: "schema",
+          provenance,
+          values: entry.values
+        })
+      )
+    );
+    candidates.relations.push(
+      ...parsed.relations.map((/** @type {any} */ relation) =>
+        makeCandidateRecord({
+          kind: "relation",
+          idHint: slugify(`${relation.from_entity}_${relation.relation_field}_${relation.to_entity}`),
+          label: `${relation.from_entity} -> ${relation.to_entity}`,
+          confidence: "high",
+          sourceKind: "schema",
+          provenance,
+          ...relation
+        })
+      )
+    );
+    candidates.indexes.push(
+      ...parsed.indexes.map((/** @type {any} */ index) =>
+        makeCandidateRecord({
+          kind: "index",
+          idHint: index.id_hint,
+          label: titleCase(index.id_hint.replace(/^index_/, "")),
+          confidence: "medium",
+          sourceKind: "schema",
+          provenance,
+          entity: index.entity,
+          fields: index.fields,
+          unique: index.unique
+        })
+      )
+    );
+  }
+
+  for (const filePath of sqlFiles) {
+    const parsed = parseSqlSchema(readTextIfExists(filePath) || "");
+    const provenance = relativeTo(paths.repoRoot, filePath);
+    hasPrimarySchemaSource = true;
+    findings.push({
+      kind: "sql_schema",
+      file: provenance,
+      entity_count: parsed.entities.length,
+      enum_count: parsed.enums.length
+    });
+    candidates.entities.push(
+      ...parsed.entities.map((/** @type {any} */ entity) =>
+        makeCandidateRecord({
+          kind: "entity",
+          idHint: `entity_${slugify(entity.name)}`,
+          label: titleCase(entity.name),
+          confidence: /migration/i.test(filePath) ? "medium" : "high",
+          sourceKind: /migration/i.test(filePath) ? "migration" : "schema",
+          provenance,
+          table_name: entity.table_name || slugify(entity.name),
+          fields: entity.fields
+        })
+      )
+    );
+    candidates.enums.push(
+      ...parsed.enums.map((/** @type {any} */ entry) =>
+        makeCandidateRecord({
+          kind: "enum",
+          idHint: idHintify(entry.name),
+          label: titleCase(entry.name),
+          confidence: /migration/i.test(filePath) ? "medium" : "high",
+          sourceKind: /migration/i.test(filePath) ? "migration" : "schema",
+          provenance,
+          values: entry.values
+        })
+      )
+    );
+    candidates.relations.push(
+      ...parsed.relations.map((/** @type {any} */ relation) =>
+        makeCandidateRecord({
+          kind: "relation",
+          idHint: slugify(`${relation.from_entity}_${relation.relation_field}_${relation.to_entity}`),
+          label: `${relation.from_entity} -> ${relation.to_entity}`,
+          confidence: "medium",
+          sourceKind: /migration/i.test(filePath) ? "migration" : "schema",
+          provenance,
+          ...relation
+        })
+      )
+    );
+    candidates.indexes.push(
+      ...parsed.indexes.map((/** @type {any} */ index) =>
+        makeCandidateRecord({
+          kind: "index",
+          idHint: index.id_hint,
+          label: titleCase(index.id_hint.replace(/^index_/, "")),
+          confidence: "medium",
+          sourceKind: /migration/i.test(filePath) ? "migration" : "schema",
+          provenance,
+          entity: index.entity,
+          fields: index.fields,
+          unique: index.unique
+        })
+      )
+    );
+  }
+
+  if (!hasPrimarySchemaSource) {
+    for (const filePath of snapshotFiles) {
+      const snapshot = readJsonIfExists(filePath);
+      if (!snapshot) {
+        continue;
+      }
+      const parsed = parseDbSchemaSnapshot(snapshot);
+      const provenance = relativeTo(paths.repoRoot, filePath);
+      findings.push({
+        kind: "db_schema_snapshot",
+        file: provenance,
+        entity_count: parsed.entities.length,
+        enum_count: parsed.enums.length
+      });
+      candidates.entities.push(
+        ...parsed.entities.map((/** @type {any} */ entity) =>
+          makeCandidateRecord({
+            kind: "entity",
+            idHint: `entity_${slugify(entity.name)}`,
+            label: titleCase(entity.name),
+            confidence: "medium",
+            sourceKind: "generated_artifact",
+            provenance,
+            table_name: entity.table_name || slugify(entity.name),
+            fields: entity.fields
+          })
+        )
+      );
+      candidates.enums.push(
+        ...parsed.enums.map((/** @type {any} */ entry) =>
+          makeCandidateRecord({
+            kind: "enum",
+            idHint: idHintify(entry.name),
+            label: titleCase(entry.name),
+            confidence: "medium",
+            sourceKind: "generated_artifact",
+            provenance,
+            values: entry.values
+          })
+        )
+      );
+      candidates.relations.push(
+        ...parsed.relations.map((/** @type {any} */ relation) =>
+          makeCandidateRecord({
+            kind: "relation",
+            idHint: slugify(`${relation.from_entity}_${relation.relation_field}_${relation.to_entity}`),
+            label: `${relation.from_entity} -> ${relation.to_entity}`,
+            confidence: "medium",
+            sourceKind: "generated_artifact",
+            provenance,
+            ...relation
+          })
+        )
+      );
+      candidates.indexes.push(
+        ...parsed.indexes.map((/** @type {any} */ index) =>
+          makeCandidateRecord({
+            kind: "index",
+            idHint: index.id_hint,
+            label: titleCase(index.id_hint.replace(/^index_/, "")),
+            confidence: "medium",
+            sourceKind: "generated_artifact",
+            provenance,
+            entity: index.entity,
+            fields: index.fields,
+            unique: index.unique
+          })
+        )
+      );
+    }
+  } else {
+    for (const filePath of snapshotFiles) {
+      findings.push({
+        kind: "db_schema_snapshot",
+        file: relativeTo(paths.repoRoot, filePath),
+        used_as_primary: false
+      });
+    }
+  }
+
+  candidates.entities = dedupeCandidateRecords(candidates.entities, (/** @type {any} */ record) => record.id_hint);
+  candidates.enums = dedupeCandidateRecords(candidates.enums, (/** @type {any} */ record) => record.id_hint);
+  candidates.relations = dedupeCandidateRecords(candidates.relations, (/** @type {any} */ record) => record.id_hint);
+  candidates.indexes = dedupeCandidateRecords(candidates.indexes, (/** @type {any} */ record) => record.id_hint);
+
+  return { findings, candidates };
+}

package/src/workflows/import-app/index.js

@@ -0,0 +1,30 @@
+// @ts-check
+import { runImportAppWorkflow } from "../../import/index.js";
+import { scanDocsWorkflow } from "../docs.js";
+
+export { collectApiImport, discoverApiSources } from "./api.js";
+export { collectDbImport, discoverDbSources } from "./db.js";
+export {
+  IMPORT_TRACKS,
+  SCALAR_FIELD_TYPES,
+  dedupeCandidateRecords,
+  findImportFiles,
+  importSearchRoots,
+  inferCapabilityEntityId,
+  makeCandidateRecord,
+  normalizeEndpointPathForMatch,
+  normalizeImportRelativePath,
+  normalizeOpenApiPath,
+  parseImportTracks,
+  selectPreferredImportFiles
+} from "./shared.js";
+export { collectUiImport } from "./ui.js";
+export { collectWorkflowImport } from "./workflow.js";
+
+/** @param {string} inputPath @param {WorkflowOptions} options @returns {any} */
+export function importAppWorkflow(inputPath, options = {}) {
+  return runImportAppWorkflow(inputPath, {
+    ...options,
+    scanDocsSummary: () => scanDocsWorkflow(inputPath).summary
+  });
+}