Implemented queries for tables (#729)

plugs/index/index.plug.yaml

@@ -25,11 +25,11 @@ functions:
   objectSourceProvider:
     path: api.ts:objectSourceProvider
     events:
-    - query:*
+      - query:*
   discoverSources:
     path: api.ts:discoverSources
     events:
-    - query_
+      - query_
   clearIndex:
     path: api.ts:clearIndex
     env: server
@@ -40,7 +40,7 @@ functions:
     events:
       - page:saved
       - page:deleted
-      
+
   parseIndexTextRepublish:
     path: "./command.ts:parseIndexTextRepublish"
     env: server
@@ -55,9 +55,9 @@ functions:
   processIndexQueue:
     path: ./command.ts:processIndexQueue
     mqSubscriptions:
-    - queue: indexQueue
-      batchSize: 10
-      autoAck: true
+      - queue: indexQueue
+        batchSize: 10
+        autoAck: true
   reindexSpace:
     path: "./command.ts:reindexSpace"
     env: server
@@ -70,45 +70,51 @@ functions:
   indexParagraphs:
     path: "./paragraph.ts:indexParagraphs"
     events:
-    - page:index
+      - page:index
 
   # Backlinks
   indexLinks:
     path: "./page_links.ts:indexLinks"
     events:
-    - page:index
+      - page:index
 
   attributeComplete:
     path: "./attributes.ts:attributeComplete"
     events:
-    - editor:complete
+      - editor:complete
 
   objectAttributeCompleter:
     path: ./attributes.ts:objectAttributeCompleter
     events:
-    - attribute:complete:*
+      - attribute:complete:*
 
   # Item indexing
   indexItem:
     path: "./item.ts:indexItems"
     events:
-    - page:index
+      - page:index
 
   # Anchors
   indexAnchors:
     path: "./anchor.ts:indexAnchors"
     events:
-    - page:index
+      - page:index
   anchorComplete:
     path: "./anchor.ts:anchorComplete"
     events:
       - editor:complete
 
+  # Tables
+  indexTables:
+    path: "table.ts:indexTables"
+    events:
+      - page:index
+
   # Headers
   indexHeaders:
     path: header.ts:indexHeaders
     events:
-    - page:index
+      - page:index
   headerComplete:
     path: header.ts:headerComplete
     events:
@@ -118,20 +124,19 @@ functions:
   indexData:
     path: data.ts:indexData
     events:
-    - page:index
+      - page:index
 
   # Script
   indexSpaceScript:
     path: script.ts:indexSpaceScript
     events:
-    - page:index
-
+      - page:index
 
   # Hashtags
   indexTags:
     path: tags.ts:indexTags
     events:
-    - page:index
+      - page:index
   tagComplete:
     path: tags.ts:tagComplete
     events:
@@ -181,7 +186,7 @@ functions:
   lintYAML:
     path: lint.ts:lintYAML
     events:
-    - editor:lint
+      - editor:lint
 
   # Tag file system
   readFileTag:
@@ -205,9 +210,9 @@ functions:
   syscallSourceProvider:
     path: builtins.ts:syscallSourceProvider
     events:
-    - query:syscall
+      - query:syscall
 
   commandSourceProvider:
     path: builtins.ts:commandSourceProvider
     events:
-    - query:command
+      - query:command

plugs/index/table.ts

@@ -0,0 +1,76 @@
+import { CompleteEvent, IndexTreeEvent, ObjectValue } from "$type/types.ts";
+import { collectNodesMatching, ParseTree } from "$lib/tree.ts";
+import { indexObjects } from "./api.ts";
+import { collectNodesOfType } from "$lib/tree.ts";
+
+type TableRowObject =
+  & ObjectValue<{
+    tableref: string;
+    page: string;
+    pos: number;
+  }>
+  & Record<string, any>;
+
+/**
+ * Replace any invalid characters in a string so it can serve as indexed field name in a query
+ * @param str the input string
+ * @returns All lowercase string with special chars replaced with underscore
+ */
+function cleanHeaderFieldName(str: string): string {
+  return str.replace(/[\W_]+/g, "_").toLowerCase();
+}
+
+/**
+ * Concat text properties of all child nodes
+ * @param nodes
+ * @returns
+ */
+function concatChildrenTexts(nodes: ParseTree[]): string {
+  return nodes.map((c) => c.text).join("").trim();
+}
+
+export async function indexTables({ name: pageName, tree }: IndexTreeEvent) {
+  console.log(`Indexing tables in '${pageName}'`);
+  const result: ObjectValue<TableRowObject>[] = [];
+
+  collectNodesMatching(tree, (t) => !!t.type?.startsWith("Table")).forEach(
+    (table) => {
+      const rows = collectNodesOfType(table, "TableRow");
+      const header = collectNodesOfType(table, "TableHeader")[0]; //Use first header. As per markdown spec there can only be exactly one
+      const headerLabels = collectNodesOfType(header, "TableCell").map((cell) =>
+        concatChildrenTexts(cell.children!)
+      ).map(cleanHeaderFieldName);
+      //console.log("Header labels", headerLabels);
+
+      for (const row of rows) {
+        const tags = new Set<string>();
+        collectNodesOfType(row, "Hashtag").forEach((h) => {
+          // Push tag to the list, removing the initial #
+          tags.add(h.children![0].text!.substring(1));
+        });
+
+        const cells = collectNodesOfType(row, "TableCell");
+
+        const tableRow: TableRowObject = {
+          tableref: `${pageName}@${table.from}`,
+          ref: `${pageName}@${row.from}`,
+          tag: "table",
+          tags: [...tags],
+          page: pageName,
+          pos: row.from!,
+        };
+        cells.forEach((c, i) => {
+          const content = concatChildrenTexts(c.children!);
+          const label = headerLabels[i];
+          tableRow[label!] = content;
+        });
+        result.push(tableRow);
+
+        //console.log("Cells", cells);
+      }
+    },
+  );
+
+  console.log("Found", result.length, "row(s)");
+  await indexObjects(pageName, result);
+}

website/Objects.md

@@ -37,6 +37,22 @@ Note that you can also query this page using the `example-tag` directly:
 example-tag
 ```
 
+## table
+Markdown table rows are indexed using the `table` tag, any additional tags can be added using [[Tags]] in any of its cells.
+
+| Title | Description Text |
+| --- | ----- |
+| This is some key | The value contains a #table-tag |
+| Some Row | This is an example row in between two others |
+| Another key | This time without a tag |
+
+
+```query
+table
+```
+
+Table headers will be normalized by converting them to lowercase and replacing all non alphanumeric characters with `_`.
+
 ## task
 
 task