@loaders.gl/graphs 5.0.0-alpha.9 → 5.0.0-beta.10

package/dist/dot-loader.cjs

@@ -44,32 +44,190 @@ var init_dot_loader_types = __esm({
       mimeTypes: ["text/vnd.graphviz", "text/x-graphviz", "application/vnd.graphviz"],
       text: true,
       worker: false,
+      supportsNativeArrow: true,
       dataType: null,
       batchType: null,
-      options: {},
+      options: { dot: { shape: "arrow" } },
       preload
     };
   }
 });
 
+// src/lib/build-graph-tables.ts
+function buildGraphTables(nodes, edges, shape, declarations) {
+  if (shape === "plain-graph-data") {
+    return { shape, nodes: Array.from(nodes), edges: Array.from(edges) };
+  }
+  if (shape !== "arrow" && shape !== "object-row-table") {
+    throw new Error(`Unsupported graph shape: ${shape}`);
+  }
+  const nodeTable = buildGraphTable(nodes, false, shape, declarations == null ? void 0 : declarations.nodes);
+  const edgeTable = buildGraphTable(edges, true, shape, declarations == null ? void 0 : declarations.edges);
+  return {
+    shape: "tables",
+    tables: [
+      { name: "nodes", table: nodeTable },
+      { name: "edges", table: edgeTable }
+    ]
+  };
+}
+function buildGraphTable(records, isEdge, shape, declarations) {
+  const columnNames = isEdge ? ["id", "sourceId", "targetId", "directed", "label", "attributes"] : ["id", "label", "attributes"];
+  const columns = new Map(columnNames.map((name) => [name, []]));
+  for (const record of records) {
+    for (const name of columnNames) {
+      const value = record[name];
+      columns.get(name).push(
+        name === "id" || name === "sourceId" || name === "targetId" ? String(value) : value ?? null
+      );
+    }
+  }
+  if (shape === "object-row-table") {
+    return {
+      shape,
+      data: columns.get("id").map(
+        (_, index) => Object.fromEntries(columnNames.map((name) => [name, columns.get(name)[index]]))
+      )
+    };
+  }
+  const vectors = /* @__PURE__ */ Object.create(null);
+  for (const [name, values] of columns) {
+    const type = name === "attributes" ? inferAttributeType(values, declarations) : name === "directed" ? new import_apache_arrow.Bool() : new import_apache_arrow.Utf8();
+    vectors[name] = (0, import_apache_arrow.vectorFromArray)(
+      values.map((value) => normalizeValue(value, type)),
+      type
+    );
+  }
+  return new import_apache_arrow.Table(vectors);
+}
+function inferAttributeType(values, declarations) {
+  const names = new Set(declarations == null ? void 0 : declarations.keys());
+  for (const value of values) {
+    if (value && typeof value === "object") {
+      for (const name of Object.keys(value)) {
+        names.add(name);
+      }
+    }
+  }
+  return new import_apache_arrow.Struct(
+    Array.from(names, (name) => {
+      const fieldValues = values.map(
+        (value) => (value == null ? void 0 : value[name]) ?? null
+      );
+      return new import_apache_arrow.Field(name, inferValueType(fieldValues, declarations == null ? void 0 : declarations.get(name)), true);
+    })
+  );
+}
+function inferValueType(values, declaration) {
+  const present = values.filter((value) => value !== null && value !== void 0);
+  if (declaration && typeof declaration === "object") {
+    if (!present.every((value) => Array.isArray(value))) return new import_apache_arrow.Utf8();
+    return new import_apache_arrow.List(
+      new import_apache_arrow.Field(
+        "item",
+        inferValueType(
+          present.flatMap((value) => value),
+          declaration.list
+        ),
+        true
+      )
+    );
+  }
+  if (declaration) {
+    const declaredTypes = {
+      boolean: new import_apache_arrow.Bool(),
+      int: new import_apache_arrow.Int32(),
+      long: new import_apache_arrow.Int64(),
+      float: new import_apache_arrow.Float32(),
+      double: new import_apache_arrow.Float64(),
+      string: new import_apache_arrow.Utf8()
+    };
+    const valid = present.every((value) => {
+      switch (declaration) {
+        case "boolean":
+          return typeof value === "boolean";
+        case "int":
+          return typeof value === "number" && Number.isInteger(value) && value >= -2147483648 && value <= 2147483647;
+        case "long":
+          return typeof value === "bigint" && value >= -(2n ** 63n) && value < 2n ** 63n;
+        case "float":
+        case "double":
+          return typeof value === "number";
+        case "string":
+          return typeof value === "string";
+      }
+    });
+    return valid ? declaredTypes[declaration] : new import_apache_arrow.Utf8();
+  }
+  if (present.length && present.every((value) => typeof value === "boolean")) return new import_apache_arrow.Bool();
+  if (present.length && present.every((value) => typeof value === "number")) return new import_apache_arrow.Float64();
+  if (present.length && present.every((value) => Array.isArray(value))) {
+    return new import_apache_arrow.List(
+      new import_apache_arrow.Field("item", inferValueType(present.flatMap((value) => value)), true)
+    );
+  }
+  if (present.length && present.every((value) => typeof value === "object" && !Array.isArray(value))) {
+    return inferAttributeType(present);
+  }
+  return new import_apache_arrow.Utf8();
+}
+function normalizeValue(value, type) {
+  if (value === null || value === void 0) return null;
+  if (type instanceof import_apache_arrow.Struct) {
+    return Object.fromEntries(
+      type.children.map((field) => [
+        field.name,
+        normalizeValue(value[field.name], field.type)
+      ])
+    );
+  }
+  if (type instanceof import_apache_arrow.List)
+    return value.map((item) => normalizeValue(item, type.valueType));
+  if (type instanceof import_apache_arrow.Utf8 && typeof value !== "string") {
+    return typeof value === "object" ? JSON.stringify(value) : String(value);
+  }
+  return value;
+}
+var import_apache_arrow;
+var init_build_graph_tables = __esm({
+  "src/lib/build-graph-tables.ts"() {
+    "use strict";
+    import_apache_arrow = require("apache-arrow");
+  }
+});
+
 // src/lib/parse-dot.ts
-function parseDOT(input) {
+function parseDOT(input, shape = "plain-graph-data") {
   const parsed = parseDOTDocument(input);
-  const subgraphs = Array.from(parsed.subgraphs.values(), (subgraph) => ({
-    ...subgraph,
-    attributes: { ...subgraph.attributes }
-  }));
   return {
-    shape: "plain-graph-data",
-    nodes: Array.from(parsed.nodes.values(), (node) => ({
+    ...buildGraphTables(iterateNodes(parsed), iterateEdges(parsed), shape),
+    metadata: {
+      id: parsed.id,
+      directed: parsed.directed,
+      strict: parsed.strict,
+      attributes: { ...parsed.graphAttributes },
+      subgraphs: Array.from(parsed.subgraphs.values(), (subgraph) => ({
+        ...subgraph,
+        attributes: { ...subgraph.attributes }
+      }))
+    }
+  };
+}
+function* iterateNodes(parsed) {
+  for (const node of parsed.nodes.values()) {
+    yield {
       id: node.id,
       ...typeof node.attributes.label === "string" ? { label: node.attributes.label } : {},
       attributes: {
         ...node.attributes,
         ...node.subgraphs.size ? { subgraphs: Array.from(node.subgraphs, (id) => describeSubgraph(id, parsed.subgraphs)) } : {}
       }
-    })),
-    edges: parsed.edges.map((edge) => ({
+    };
+  }
+}
+function* iterateEdges(parsed) {
+  for (const edge of parsed.edges) {
+    yield {
       id: edge.id,
       sourceId: edge.sourceId,
       targetId: edge.targetId,
@@ -79,15 +237,8 @@ function parseDOT(input) {
         ...edge.attributes,
         ...edge.subgraphs.length ? { subgraphs: edge.subgraphs.map((id) => describeSubgraph(id, parsed.subgraphs)) } : {}
       }
-    })),
-    metadata: {
-      id: parsed.id,
-      directed: parsed.directed,
-      strict: parsed.strict,
-      attributes: { ...parsed.graphAttributes },
-      subgraphs
-    }
-  };
+    };
+  }
 }
 function describeSubgraph(id, descriptors) {
   const subgraph = descriptors.get(id);
@@ -239,6 +390,10 @@ function readQuotedString(input, startIndex) {
 function readEscapedCharacter(input, startIndex) {
   const next = input[startIndex];
   switch (next) {
+    case "\n":
+      return { value: "", nextIndex: startIndex + 1 };
+    case "\r":
+      return { value: "", nextIndex: startIndex + (input[startIndex + 1] === "\n" ? 2 : 1) };
     case "n":
     case "l":
     case "L":
@@ -255,7 +410,7 @@ function readEscapedCharacter(input, startIndex) {
       if (typeof next === "undefined") {
         throw new Error("Unterminated escape sequence in DOT source.");
       }
-      return { value: next, nextIndex: startIndex + 1 };
+      return { value: `\\${next}`, nextIndex: startIndex + 1 };
     }
   }
 }
@@ -374,6 +529,7 @@ var DOTParser, IDENTIFIER_TERMINATORS;
 var init_parse_dot = __esm({
   "src/lib/parse-dot.ts"() {
     "use strict";
+    init_build_graph_tables();
     DOTParser = class {
       /** Tokens for DOT parsing. */
       tokens;
@@ -577,6 +733,10 @@ var init_parse_dot = __esm({
           const existingEdge = this.result.strict ? this.strictEdges.get(strictKey) : void 0;
           if (existingEdge) {
             existingEdge.attributes = { ...existingEdge.attributes, ...explicitAttributes };
+            const explicitId = explicitAttributes.id ?? explicitAttributes.Id ?? explicitAttributes.ID;
+            if (typeof explicitId === "string" || typeof explicitId === "number") {
+              existingEdge.id = String(explicitId);
+            }
             existingEdge.directed = deriveDirectedFlag(existingEdge.attributes, directed);
             existingEdge.subgraphs = Array.from(/* @__PURE__ */ new Set([...existingEdge.subgraphs, ...membership]));
             continue;
@@ -755,9 +915,18 @@ var init_dot_loader = __esm({
     init_parse_dot();
     DOTLoaderWithParser = {
       ...DOTLoader,
-      parse: async (data) => parseDOT(new TextDecoder().decode(data)),
-      parseSync: (data) => parseDOT(new TextDecoder().decode(data)),
-      parseTextSync: parseDOT
+      parse: async (data, options) => {
+        var _a;
+        return parseDOT(new TextDecoder().decode(data), ((_a = options == null ? void 0 : options.dot) == null ? void 0 : _a.shape) ?? "arrow");
+      },
+      parseSync: (data, options) => {
+        var _a;
+        return parseDOT(new TextDecoder().decode(data), ((_a = options == null ? void 0 : options.dot) == null ? void 0 : _a.shape) ?? "arrow");
+      },
+      parseTextSync: (data, options) => {
+        var _a;
+        return parseDOT(data, ((_a = options == null ? void 0 : options.dot) == null ? void 0 : _a.shape) ?? "arrow");
+      }
     };
   }
 });
@@ -770,7 +939,7 @@ __export(dot_loader_exports2, {
 module.exports = __toCommonJS(dot_loader_exports2);
 
 // dist/dot-loader-types.js
-var VERSION2 = true ? "5.0.0-alpha.9" : "latest";
+var VERSION2 = true ? "5.0.0-beta.10" : "latest";
 async function preload2() {
   const { DOTLoaderWithParser: DOTLoaderWithParser3 } = await Promise.resolve().then(() => (init_dot_loader(), dot_loader_exports));
   return DOTLoaderWithParser3;
@@ -784,30 +953,162 @@ var DOTLoader2 = {
   mimeTypes: ["text/vnd.graphviz", "text/x-graphviz", "application/vnd.graphviz"],
   text: true,
   worker: false,
+  supportsNativeArrow: true,
   dataType: null,
   batchType: null,
-  options: {},
+  options: { dot: { shape: "arrow" } },
   preload: preload2
 };
 
+// dist/lib/build-graph-tables.js
+var import_apache_arrow2 = require("apache-arrow");
+function buildGraphTables2(nodes, edges, shape, declarations) {
+  if (shape === "plain-graph-data") {
+    return { shape, nodes: Array.from(nodes), edges: Array.from(edges) };
+  }
+  if (shape !== "arrow" && shape !== "object-row-table") {
+    throw new Error(`Unsupported graph shape: ${shape}`);
+  }
+  const nodeTable = buildGraphTable2(nodes, false, shape, declarations == null ? void 0 : declarations.nodes);
+  const edgeTable = buildGraphTable2(edges, true, shape, declarations == null ? void 0 : declarations.edges);
+  return {
+    shape: "tables",
+    tables: [
+      { name: "nodes", table: nodeTable },
+      { name: "edges", table: edgeTable }
+    ]
+  };
+}
+function buildGraphTable2(records, isEdge, shape, declarations) {
+  const columnNames = isEdge ? ["id", "sourceId", "targetId", "directed", "label", "attributes"] : ["id", "label", "attributes"];
+  const columns = new Map(columnNames.map((name) => [name, []]));
+  for (const record of records) {
+    for (const name of columnNames) {
+      const value = record[name];
+      columns.get(name).push(name === "id" || name === "sourceId" || name === "targetId" ? String(value) : value ?? null);
+    }
+  }
+  if (shape === "object-row-table") {
+    return {
+      shape,
+      data: columns.get("id").map((_, index) => Object.fromEntries(columnNames.map((name) => [name, columns.get(name)[index]])))
+    };
+  }
+  const vectors = /* @__PURE__ */ Object.create(null);
+  for (const [name, values] of columns) {
+    const type = name === "attributes" ? inferAttributeType2(values, declarations) : name === "directed" ? new import_apache_arrow2.Bool() : new import_apache_arrow2.Utf8();
+    vectors[name] = (0, import_apache_arrow2.vectorFromArray)(values.map((value) => normalizeValue2(value, type)), type);
+  }
+  return new import_apache_arrow2.Table(vectors);
+}
+function inferAttributeType2(values, declarations) {
+  const names = new Set(declarations == null ? void 0 : declarations.keys());
+  for (const value of values) {
+    if (value && typeof value === "object") {
+      for (const name of Object.keys(value)) {
+        names.add(name);
+      }
+    }
+  }
+  return new import_apache_arrow2.Struct(Array.from(names, (name) => {
+    const fieldValues = values.map((value) => (value == null ? void 0 : value[name]) ?? null);
+    return new import_apache_arrow2.Field(name, inferValueType2(fieldValues, declarations == null ? void 0 : declarations.get(name)), true);
+  }));
+}
+function inferValueType2(values, declaration) {
+  const present = values.filter((value) => value !== null && value !== void 0);
+  if (declaration && typeof declaration === "object") {
+    if (!present.every((value) => Array.isArray(value)))
+      return new import_apache_arrow2.Utf8();
+    return new import_apache_arrow2.List(new import_apache_arrow2.Field("item", inferValueType2(present.flatMap((value) => value), declaration.list), true));
+  }
+  if (declaration) {
+    const declaredTypes = {
+      boolean: new import_apache_arrow2.Bool(),
+      int: new import_apache_arrow2.Int32(),
+      long: new import_apache_arrow2.Int64(),
+      float: new import_apache_arrow2.Float32(),
+      double: new import_apache_arrow2.Float64(),
+      string: new import_apache_arrow2.Utf8()
+    };
+    const valid = present.every((value) => {
+      switch (declaration) {
+        case "boolean":
+          return typeof value === "boolean";
+        case "int":
+          return typeof value === "number" && Number.isInteger(value) && value >= -2147483648 && value <= 2147483647;
+        case "long":
+          return typeof value === "bigint" && value >= -(2n ** 63n) && value < 2n ** 63n;
+        case "float":
+        case "double":
+          return typeof value === "number";
+        case "string":
+          return typeof value === "string";
+      }
+    });
+    return valid ? declaredTypes[declaration] : new import_apache_arrow2.Utf8();
+  }
+  if (present.length && present.every((value) => typeof value === "boolean"))
+    return new import_apache_arrow2.Bool();
+  if (present.length && present.every((value) => typeof value === "number"))
+    return new import_apache_arrow2.Float64();
+  if (present.length && present.every((value) => Array.isArray(value))) {
+    return new import_apache_arrow2.List(new import_apache_arrow2.Field("item", inferValueType2(present.flatMap((value) => value)), true));
+  }
+  if (present.length && present.every((value) => typeof value === "object" && !Array.isArray(value))) {
+    return inferAttributeType2(present);
+  }
+  return new import_apache_arrow2.Utf8();
+}
+function normalizeValue2(value, type) {
+  if (value === null || value === void 0)
+    return null;
+  if (type instanceof import_apache_arrow2.Struct) {
+    return Object.fromEntries(type.children.map((field) => [
+      field.name,
+      normalizeValue2(value[field.name], field.type)
+    ]));
+  }
+  if (type instanceof import_apache_arrow2.List)
+    return value.map((item) => normalizeValue2(item, type.valueType));
+  if (type instanceof import_apache_arrow2.Utf8 && typeof value !== "string") {
+    return typeof value === "object" ? JSON.stringify(value) : String(value);
+  }
+  return value;
+}
+
 // dist/lib/parse-dot.js
-function parseDOT2(input) {
+function parseDOT2(input, shape = "plain-graph-data") {
   const parsed = parseDOTDocument2(input);
-  const subgraphs = Array.from(parsed.subgraphs.values(), (subgraph) => ({
-    ...subgraph,
-    attributes: { ...subgraph.attributes }
-  }));
   return {
-    shape: "plain-graph-data",
-    nodes: Array.from(parsed.nodes.values(), (node) => ({
+    ...buildGraphTables2(iterateNodes2(parsed), iterateEdges2(parsed), shape),
+    metadata: {
+      id: parsed.id,
+      directed: parsed.directed,
+      strict: parsed.strict,
+      attributes: { ...parsed.graphAttributes },
+      subgraphs: Array.from(parsed.subgraphs.values(), (subgraph) => ({
+        ...subgraph,
+        attributes: { ...subgraph.attributes }
+      }))
+    }
+  };
+}
+function* iterateNodes2(parsed) {
+  for (const node of parsed.nodes.values()) {
+    yield {
       id: node.id,
       ...typeof node.attributes.label === "string" ? { label: node.attributes.label } : {},
       attributes: {
         ...node.attributes,
         ...node.subgraphs.size ? { subgraphs: Array.from(node.subgraphs, (id) => describeSubgraph2(id, parsed.subgraphs)) } : {}
       }
-    })),
-    edges: parsed.edges.map((edge) => ({
+    };
+  }
+}
+function* iterateEdges2(parsed) {
+  for (const edge of parsed.edges) {
+    yield {
       id: edge.id,
       sourceId: edge.sourceId,
       targetId: edge.targetId,
@@ -817,15 +1118,8 @@ function parseDOT2(input) {
         ...edge.attributes,
         ...edge.subgraphs.length ? { subgraphs: edge.subgraphs.map((id) => describeSubgraph2(id, parsed.subgraphs)) } : {}
       }
-    })),
-    metadata: {
-      id: parsed.id,
-      directed: parsed.directed,
-      strict: parsed.strict,
-      attributes: { ...parsed.graphAttributes },
-      subgraphs
-    }
-  };
+    };
+  }
 }
 function describeSubgraph2(id, descriptors) {
   const subgraph = descriptors.get(id);
@@ -1036,6 +1330,10 @@ var DOTParser2 = class {
       const existingEdge = this.result.strict ? this.strictEdges.get(strictKey) : void 0;
       if (existingEdge) {
         existingEdge.attributes = { ...existingEdge.attributes, ...explicitAttributes };
+        const explicitId = explicitAttributes.id ?? explicitAttributes.Id ?? explicitAttributes.ID;
+        if (typeof explicitId === "string" || typeof explicitId === "number") {
+          existingEdge.id = String(explicitId);
+        }
         existingEdge.directed = deriveDirectedFlag2(existingEdge.attributes, directed);
         existingEdge.subgraphs = Array.from(/* @__PURE__ */ new Set([...existingEdge.subgraphs, ...membership]));
         continue;
@@ -1340,6 +1638,10 @@ function readQuotedString2(input, startIndex) {
 function readEscapedCharacter2(input, startIndex) {
   const next = input[startIndex];
   switch (next) {
+    case "\n":
+      return { value: "", nextIndex: startIndex + 1 };
+    case "\r":
+      return { value: "", nextIndex: startIndex + (input[startIndex + 1] === "\n" ? 2 : 1) };
     case "n":
     case "l":
     case "L":
@@ -1356,7 +1658,7 @@ function readEscapedCharacter2(input, startIndex) {
       if (typeof next === "undefined") {
         throw new Error("Unterminated escape sequence in DOT source.");
       }
-      return { value: next, nextIndex: startIndex + 1 };
+      return { value: `\\${next}`, nextIndex: startIndex + 1 };
     }
   }
 }
@@ -1475,8 +1777,17 @@ function deriveDirectedFlag2(attributes, defaultDirected) {
 // dist/dot-loader.js
 var DOTLoaderWithParser2 = {
   ...DOTLoader2,
-  parse: async (data) => parseDOT2(new TextDecoder().decode(data)),
-  parseSync: (data) => parseDOT2(new TextDecoder().decode(data)),
-  parseTextSync: parseDOT2
+  parse: async (data, options) => {
+    var _a;
+    return parseDOT2(new TextDecoder().decode(data), ((_a = options == null ? void 0 : options.dot) == null ? void 0 : _a.shape) ?? "arrow");
+  },
+  parseSync: (data, options) => {
+    var _a;
+    return parseDOT2(new TextDecoder().decode(data), ((_a = options == null ? void 0 : options.dot) == null ? void 0 : _a.shape) ?? "arrow");
+  },
+  parseTextSync: (data, options) => {
+    var _a;
+    return parseDOT2(data, ((_a = options == null ? void 0 : options.dot) == null ? void 0 : _a.shape) ?? "arrow");
+  }
 };
 //# sourceMappingURL=dot-loader.cjs.map