fix: handle duplicate columns and tables with no header

package.json

@@ -36,6 +36,7 @@
     "eslint-config-prettier": "8.8.0",
     "eslint-plugin-jest": "27.2.1",
     "eslint-plugin-prettier": "4.2.1",
+    "hexoid": "^1.0.0",
     "husky": "8.0.3",
     "jest": "29.7.0",
     "jest-environment-jsdom": "29.7.0",

packages/@atjson/offset-annotations/src/annotations/table.ts

@@ -30,8 +30,26 @@ export class Table extends BlockAnnotation<{
    * them from rendering.
    */
   columns: Array<{
+    /**
+     * `name` corresponds to the fields on the schema of the corresponding DataSet
+     */
     name: string;
+
+    /**
+     * `plaintextName`, if present, is a stringified version of the displayed column header text
+     */
+    plaintextName?: string;
+
+    /**
+     * `slice` is a reference to the slice of the document to be rendered as the column header
+     */
     slice?: string;
+
+    /**
+     * `textAlign` is the intended alignment of all text in this column
+     *   This corresponds to the behavior of tables in markdown, where
+     *   text alignment can only be set at the column level
+     */
     textAlign?: "left" | "right" | "center";
   }>;
 

packages/@atjson/offset-annotations/src/utils/convert-html-tables.ts

@@ -1,3 +1,4 @@
+import hexoid from "hexoid";
 import {
   JSON,
   SliceAnnotation,
@@ -8,6 +9,8 @@ import {
 
 import OffsetSource, { DataSet, Table, ColumnType } from "../index";
 
+const generateId = hexoid(8);
+
 function extractPlainContents(
   doc: OffsetSource,
   annotation: { start: number; end: number }
@@ -80,15 +83,27 @@ function extractAlignment(tableCell: Annotation<{ style: string }>) {
   return undefined;
 }
 
+function generateFieldName(
+  columnName: string | null,
+  index: number,
+  tableIdentifier: string
+) {
+  return (
+    (columnName?.length ? columnName : `column ${index + 1}`) +
+    `(${tableIdentifier}-${generateId()})`
+  );
+}
+
 export function convertHTMLTablesToDataSet(
   doc: OffsetSource,
   vendor: string
-): void {
+): Table[] {
+  let convertedTables: Table[] = [];
   doc.where({ type: `-${vendor}-table` }).forEach((table) => {
     if (isInvalidTable(doc, table, vendor)) {
       return;
     }
-
+    let tableIdentifier = generateId();
     let dataSetSchemaEntries: [name: string, type: ColumnType][] = [];
     let columnConfigs: Exclude<Table["attributes"]["columns"], undefined> = [];
     let dataRows: Record<string, { slice: string; jsonValue: JSON }>[] = [];
@@ -120,16 +135,15 @@ export function convertHTMLTablesToDataSet(
         });
         doc.replaceAnnotation(headCell, slice);
         let columnName = extractPlainContents(doc, slice);
+        let fieldName = generateFieldName(columnName, index, tableIdentifier);
 
-        dataSetSchemaEntries.push([
-          columnName.length ? columnName : `column ${index + 1}`,
-          ColumnType.RICH_TEXT,
-        ]);
+        dataSetSchemaEntries.push([fieldName, ColumnType.RICH_TEXT]);
 
         slices.push(slice);
 
         let columnConfig: (typeof columnConfigs)[number] = {
-          name: columnName,
+          name: fieldName,
+          plaintextName: columnName,
           slice: slice.id,
         };
 
@@ -174,7 +188,7 @@ export function convertHTMLTablesToDataSet(
           }
 
           if (!hasColumnHeaders) {
-            let columnName = `column ${index + 1}`;
+            let columnName = generateFieldName(null, index, tableIdentifier);
 
             let columnConfig: (typeof columnConfigs)[number] = {
               name: columnName,
@@ -234,8 +248,12 @@ export function convertHTMLTablesToDataSet(
     slices.forEach((slice) => slice.attributes.refs.push(dataSet.id));
 
     doc.replaceAnnotation(table, dataSet, offsetTable);
+
+    convertedTables.push(offsetTable);
   });
 
   doc.where({ type: `-${vendor}-thead` }).remove();
   doc.where({ type: `-${vendor}-tbody` }).remove();
+
+  return convertedTables;
 }

packages/@atjson/source-commonmark/src/converter/index.ts

@@ -93,7 +93,20 @@ CommonmarkSource.defineConverterTo(
       .set({ type: "-offset-paragraph" });
     doc.where({ type: "-commonmark-strong" }).set({ type: "-offset-bold" });
 
-    convertHTMLTablesToDataSet(doc, "commonmark");
+    convertHTMLTablesToDataSet(doc, "commonmark").forEach((table) => {
+      /**
+       * if a markdown table has no text in its column headers, we have decided to
+       * store that as a table with no header section at all
+       */
+      if (
+        table.attributes.columns.every(
+          (column) =>
+            column.plaintextName === "" || column.plaintextName == null
+        )
+      ) {
+        table.attributes.showColumnHeaders = false;
+      }
+    });
 
     return doc;
   }