@rejot-dev/thalo 0.0.0

package/dist/services/semantic-tokens.d.ts

@@ -0,0 +1,54 @@
+import { ParsedDocument } from "../parser.shared.js";
+
+//#region src/services/semantic-tokens.d.ts
+
+/**
+ * Semantic token types - these map to LSP's SemanticTokenTypes
+ *
+ * See: https://microsoft.github.io/language-server-protocol/specifications/lsp/3.17/specification/#semanticTokenTypes
+ */
+declare const tokenTypes: readonly ["namespace", "type", "class", "function", "property", "string", "keyword", "comment", "variable", "number", "operator", "macro"];
+type TokenType = (typeof tokenTypes)[number];
+/**
+ * Semantic token modifiers
+ */
+declare const tokenModifiers: readonly ["declaration", "definition", "documentation", "readonly"];
+type TokenModifier = (typeof tokenModifiers)[number];
+/**
+ * A semantic token with position and type information
+ */
+interface SemanticToken {
+  /** Line number (0-based) */
+  line: number;
+  /** Character offset on the line (0-based) */
+  startChar: number;
+  /** Length of the token */
+  length: number;
+  /** Token type index (into tokenTypes array) */
+  tokenType: number;
+  /** Token modifiers as a bitmask */
+  tokenModifiers: number;
+}
+/**
+ * Get the index of a token type
+ */
+declare function getTokenTypeIndex(type: TokenType): number;
+/**
+ * Get the modifier bitmask for a set of modifiers
+ */
+declare function getTokenModifiersMask(modifiers: TokenModifier[]): number;
+/**
+ * Extract semantic tokens from a parsed document.
+ * Returns tokens with file-absolute positions (sourceMap applied).
+ */
+declare function extractSemanticTokens(document: ParsedDocument): SemanticToken[];
+/**
+ * Encode semantic tokens into the LSP delta format
+ *
+ * LSP expects tokens in a flattened array where each token is:
+ * [deltaLine, deltaStartChar, length, tokenType, tokenModifiers]
+ */
+declare function encodeSemanticTokens(tokens: SemanticToken[]): number[];
+//#endregion
+export { SemanticToken, TokenModifier, TokenType, encodeSemanticTokens, extractSemanticTokens, getTokenModifiersMask, getTokenTypeIndex, tokenModifiers, tokenTypes };
+//# sourceMappingURL=semantic-tokens.d.ts.map

package/dist/services/semantic-tokens.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"semantic-tokens.d.ts","names":[],"sources":["../../src/services/semantic-tokens.ts"],"sourcesContent":[],"mappings":";;;;;;AASA;AAeA;AAKA;AAOY,cA3BC,UA2BY,EAAW,SAAA,CAAA,WAAc,EAAA,MAAA,EAAA,OAAA,EAAA,UAAA,EAAA,UAAA,EAAA,QAAA,EAAA,SAAA,EAAA,SAAA,EAAA,UAAA,EAAA,QAAA,EAAA,UAAA,EAAA,OAAA,CAAA;AAKjC,KAjBL,SAAA,GAiBkB,CAAA,OAjBE,UAiBF,CAAA,CAAA,MAAA,CAAA;AAgB9B;AAOA;AAeA;AA+OgB,cAjSH,cAiSuB,EAAA,SAAS,CAAA,aAAa,EAAA,YAAA,EAAA,eAAA,EAAA,UAAA,CAAA;KA1R9C,aAAA,WAAwB;;;;UAKnB,aAAA;;;;;;;;;;;;;;;iBAgBD,iBAAA,OAAwB;;;;iBAOxB,qBAAA,YAAiC;;;;;iBAejC,qBAAA,WAAgC,iBAAiB;;;;;;;iBA+OjD,oBAAA,SAA6B"}

package/dist/services/semantic-tokens.js

@@ -0,0 +1,213 @@
+//#region src/services/semantic-tokens.ts
+/**
+* Semantic token types - these map to LSP's SemanticTokenTypes
+*
+* See: https://microsoft.github.io/language-server-protocol/specifications/lsp/3.17/specification/#semanticTokenTypes
+*/
+const tokenTypes = [
+	"namespace",
+	"type",
+	"class",
+	"function",
+	"property",
+	"string",
+	"keyword",
+	"comment",
+	"variable",
+	"number",
+	"operator",
+	"macro"
+];
+/**
+* Semantic token modifiers
+*/
+const tokenModifiers = [
+	"declaration",
+	"definition",
+	"documentation",
+	"readonly"
+];
+/**
+* Get the index of a token type
+*/
+function getTokenTypeIndex(type) {
+	return tokenTypes.indexOf(type);
+}
+/**
+* Get the modifier bitmask for a set of modifiers
+*/
+function getTokenModifiersMask(modifiers) {
+	let mask = 0;
+	for (const mod of modifiers) {
+		const index = tokenModifiers.indexOf(mod);
+		if (index >= 0) mask |= 1 << index;
+	}
+	return mask;
+}
+/**
+* Extract semantic tokens from a parsed document.
+* Returns tokens with file-absolute positions (sourceMap applied).
+*/
+function extractSemanticTokens(document) {
+	const tokens = [];
+	for (const block of document.blocks) extractTokensFromTree(block.tree, block.sourceMap, tokens);
+	tokens.sort((a, b) => {
+		if (a.line !== b.line) return a.line - b.line;
+		return a.startChar - b.startChar;
+	});
+	return tokens;
+}
+/**
+* Extract tokens from a tree-sitter tree
+*/
+function extractTokensFromTree(tree, sourceMap, tokens) {
+	const cursor = tree.walk();
+	const visitNode = (node) => {
+		const token = getTokenForNode(node, sourceMap);
+		if (token) tokens.push(token);
+		if (cursor.gotoFirstChild()) {
+			do
+				visitNode(cursor.currentNode);
+			while (cursor.gotoNextSibling());
+			cursor.gotoParent();
+		}
+	};
+	visitNode(cursor.currentNode);
+}
+/**
+* Get a semantic token for a tree-sitter node (if applicable).
+* Returns token with file-absolute positions (sourceMap applied).
+*/
+function getTokenForNode(node, sourceMap) {
+	let { type, startPosition, endPosition } = node;
+	let nodeText = node.text;
+	if (type === "ERROR") return null;
+	let tokenType = null;
+	let modifiers = [];
+	switch (type) {
+		case "timestamp":
+			tokenType = "number";
+			break;
+		case "instance_directive":
+		case "schema_directive":
+		case "define-synthesis":
+		case "actualize-synthesis":
+			tokenType = "keyword";
+			break;
+		case "entity":
+		case "query_entity":
+			tokenType = "type";
+			break;
+		case "identifier":
+			tokenType = "type";
+			modifiers = ["definition"];
+			break;
+		case "title":
+			tokenType = "string";
+			break;
+		case "link":
+			tokenType = "function";
+			if (node.parent?.type === "data_entry" || node.parent?.type === "schema_entry") modifiers = ["declaration"];
+			break;
+		case "tag":
+			tokenType = "variable";
+			break;
+		case "key":
+			tokenType = "property";
+			break;
+		case "quoted_value":
+			tokenType = "string";
+			break;
+		case "datetime_value":
+		case "date_range":
+			tokenType = "number";
+			break;
+		case "field_name":
+		case "section_name": {
+			tokenType = "property";
+			const trimmed = nodeText.trimStart();
+			if (nodeText.length - trimmed.length > 0) {
+				startPosition = {
+					row: endPosition.row,
+					column: endPosition.column - trimmed.length
+				};
+				nodeText = trimmed;
+			}
+			break;
+		}
+		case "optional_marker":
+			tokenType = "operator";
+			modifiers = ["readonly"];
+			break;
+		case "primitive_type":
+			tokenType = "type";
+			break;
+		case "literal_type":
+			tokenType = "string";
+			break;
+		case "description":
+			tokenType = "string";
+			modifiers = ["documentation"];
+			break;
+		case "|":
+		case "[]":
+		case ":":
+		case "=":
+		case ";":
+		case "(":
+		case ")":
+		case ",":
+			tokenType = "operator";
+			break;
+		case "where":
+		case "and":
+			tokenType = "keyword";
+			break;
+		case "condition_field":
+			tokenType = "property";
+			break;
+		case "comment":
+			tokenType = "comment";
+			break;
+		case "md_indicator":
+			tokenType = "macro";
+			break;
+		case "md_heading_text":
+			tokenType = "keyword";
+			break;
+		default: return null;
+	}
+	if (!tokenType) return null;
+	const length = startPosition.row === endPosition.row ? endPosition.column - startPosition.column : nodeText.indexOf("\n");
+	const isFirstBlockLine = startPosition.row === 0;
+	return {
+		line: sourceMap.lineOffset + startPosition.row,
+		startChar: isFirstBlockLine ? sourceMap.columnOffset + startPosition.column : startPosition.column,
+		length: length > 0 ? length : nodeText.length,
+		tokenType: getTokenTypeIndex(tokenType),
+		tokenModifiers: getTokenModifiersMask(modifiers)
+	};
+}
+/**
+* Encode semantic tokens into the LSP delta format
+*
+* LSP expects tokens in a flattened array where each token is:
+* [deltaLine, deltaStartChar, length, tokenType, tokenModifiers]
+*/
+function encodeSemanticTokens(tokens) {
+	const data = [];
+	let prevLine = 0;
+	let prevChar = 0;
+	for (const token of tokens) {
+		const deltaLine = token.line - prevLine;
+		const deltaChar = deltaLine === 0 ? token.startChar - prevChar : token.startChar;
+		data.push(deltaLine, deltaChar, token.length, token.tokenType, token.tokenModifiers);
+		prevLine = token.line;
+		prevChar = token.startChar;
+	}
+	return data;
+}
+
+//#endregion
+export { encodeSemanticTokens, extractSemanticTokens, getTokenModifiersMask, getTokenTypeIndex, tokenModifiers, tokenTypes };
+//# sourceMappingURL=semantic-tokens.js.map

package/dist/services/semantic-tokens.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"semantic-tokens.js","names":["tokens: SemanticToken[]","tokenType: TokenType | null","modifiers: TokenModifier[]","data: number[]"],"sources":["../../src/services/semantic-tokens.ts"],"sourcesContent":["import type { SyntaxNode } from \"../ast/ast-types.js\";\nimport type { ParsedDocument, GenericTree } from \"../parser.shared.js\";\nimport type { SourceMap } from \"../source-map.js\";\n\n/**\n * Semantic token types - these map to LSP's SemanticTokenTypes\n *\n * See: https://microsoft.github.io/language-server-protocol/specifications/lsp/3.17/specification/#semanticTokenTypes\n */\nexport const tokenTypes = [\n  \"namespace\", // timestamp\n  \"type\", // entity type (lore, opinion, etc.), primitive types\n  \"class\", // (unused, kept for index stability)\n  \"function\", // link (^link-id)\n  \"property\", // metadata key, field name, section name\n  \"string\", // title, description, quoted strings, literal types\n  \"keyword\", // where, and, block headers (# Metadata, etc.)\n  \"comment\", // comments\n  \"variable\", // tag names\n  \"number\", // datetime values, date ranges\n  \"operator\", // |, [], =, :, ;, (, ), ,\n  \"macro\", // markdown header indicator (#, ##, etc.)\n] as const;\n\nexport type TokenType = (typeof tokenTypes)[number];\n\n/**\n * Semantic token modifiers\n */\nexport const tokenModifiers = [\n  \"declaration\", // link definition\n  \"definition\", // entity definition\n  \"documentation\", // description text\n  \"readonly\", // optional marker (?)\n] as const;\n\nexport type TokenModifier = (typeof tokenModifiers)[number];\n\n/**\n * A semantic token with position and type information\n */\nexport interface SemanticToken {\n  /** Line number (0-based) */\n  line: number;\n  /** Character offset on the line (0-based) */\n  startChar: number;\n  /** Length of the token */\n  length: number;\n  /** Token type index (into tokenTypes array) */\n  tokenType: number;\n  /** Token modifiers as a bitmask */\n  tokenModifiers: number;\n}\n\n/**\n * Get the index of a token type\n */\nexport function getTokenTypeIndex(type: TokenType): number {\n  return tokenTypes.indexOf(type);\n}\n\n/**\n * Get the modifier bitmask for a set of modifiers\n */\nexport function getTokenModifiersMask(modifiers: TokenModifier[]): number {\n  let mask = 0;\n  for (const mod of modifiers) {\n    const index = tokenModifiers.indexOf(mod);\n    if (index >= 0) {\n      mask |= 1 << index;\n    }\n  }\n  return mask;\n}\n\n/**\n * Extract semantic tokens from a parsed document.\n * Returns tokens with file-absolute positions (sourceMap applied).\n */\nexport function extractSemanticTokens(document: ParsedDocument): SemanticToken[] {\n  const tokens: SemanticToken[] = [];\n\n  for (const block of document.blocks) {\n    extractTokensFromTree(block.tree, block.sourceMap, tokens);\n  }\n\n  // Sort by position (line, then character)\n  tokens.sort((a, b) => {\n    if (a.line !== b.line) {\n      return a.line - b.line;\n    }\n    return a.startChar - b.startChar;\n  });\n\n  return tokens;\n}\n\n/**\n * Extract tokens from a tree-sitter tree\n */\nfunction extractTokensFromTree(\n  tree: GenericTree,\n  sourceMap: SourceMap,\n  tokens: SemanticToken[],\n): void {\n  // eslint-disable-next-line @typescript-eslint/no-explicit-any\n  const cursor = (tree as any).walk();\n\n  // DFS traversal\n  const visitNode = (node: SyntaxNode): void => {\n    const token = getTokenForNode(node, sourceMap);\n    if (token) {\n      tokens.push(token);\n    }\n\n    // Visit children\n    if (cursor.gotoFirstChild()) {\n      do {\n        visitNode(cursor.currentNode);\n      } while (cursor.gotoNextSibling());\n      cursor.gotoParent();\n    }\n  };\n\n  visitNode(cursor.currentNode);\n}\n\n/**\n * Get a semantic token for a tree-sitter node (if applicable).\n * Returns token with file-absolute positions (sourceMap applied).\n */\nfunction getTokenForNode(node: SyntaxNode, sourceMap: SourceMap): SemanticToken | null {\n  let { type, startPosition, endPosition } = node;\n  let nodeText = node.text;\n\n  // Skip ERROR nodes and certain wrapper nodes\n  if (type === \"ERROR\") {\n    return null;\n  }\n\n  let tokenType: TokenType | null = null;\n  let modifiers: TokenModifier[] = [];\n\n  switch (type) {\n    // =========================================================================\n    // Header elements\n    // =========================================================================\n    case \"timestamp\":\n      tokenType = \"number\"; // same as datetime_value for consistent date coloring\n      break;\n\n    case \"instance_directive\":\n    case \"schema_directive\":\n    case \"define-synthesis\":\n    case \"actualize-synthesis\":\n      tokenType = \"keyword\";\n      break;\n\n    case \"entity\":\n    case \"query_entity\":\n      tokenType = \"type\";\n      break;\n\n    case \"identifier\":\n      // Entity name in schema entries (define-entity custom)\n      tokenType = \"type\";\n      modifiers = [\"definition\"];\n      break;\n\n    case \"title\":\n      tokenType = \"string\";\n      break;\n\n    case \"link\":\n      tokenType = \"function\";\n      // Check if this is a definition (directly on entry) or reference (in metadata)\n      if (node.parent?.type === \"data_entry\" || node.parent?.type === \"schema_entry\") {\n        modifiers = [\"declaration\"];\n      }\n      break;\n\n    case \"tag\":\n      tokenType = \"variable\";\n      break;\n\n    // =========================================================================\n    // Metadata\n    // =========================================================================\n    case \"key\":\n      tokenType = \"property\";\n      break;\n\n    case \"quoted_value\":\n      tokenType = \"string\";\n      break;\n\n    case \"datetime_value\":\n    case \"date_range\":\n      tokenType = \"number\";\n      break;\n\n    // =========================================================================\n    // Schema definitions\n    // =========================================================================\n    case \"field_name\":\n    case \"section_name\": {\n      tokenType = \"property\";\n      // These tokens include leading whitespace (\\n + indent), strip it\n      const trimmed = nodeText.trimStart();\n      const leadingLen = nodeText.length - trimmed.length;\n      if (leadingLen > 0) {\n        // Adjust position to after the whitespace (endPosition is correct, work backwards)\n        startPosition = { row: endPosition.row, column: endPosition.column - trimmed.length };\n        nodeText = trimmed;\n      }\n      break;\n    }\n\n    case \"optional_marker\":\n      tokenType = \"operator\";\n      modifiers = [\"readonly\"];\n      break;\n\n    case \"primitive_type\":\n      tokenType = \"type\";\n      break;\n\n    case \"literal_type\":\n      tokenType = \"string\";\n      break;\n\n    case \"description\":\n      tokenType = \"string\";\n      modifiers = [\"documentation\"];\n      break;\n\n    // =========================================================================\n    // Operators and punctuation\n    // =========================================================================\n    case \"|\":\n    case \"[]\":\n    case \":\":\n    case \"=\":\n    case \";\":\n    case \"(\":\n    case \")\":\n    case \",\":\n      tokenType = \"operator\";\n      break;\n\n    // =========================================================================\n    // Query expressions\n    // =========================================================================\n    case \"where\":\n    case \"and\":\n      tokenType = \"keyword\";\n      break;\n\n    case \"condition_field\":\n      tokenType = \"property\";\n      break;\n\n    // =========================================================================\n    // Comments\n    // =========================================================================\n    case \"comment\":\n      tokenType = \"comment\";\n      break;\n\n    // =========================================================================\n    // Content sections\n    // =========================================================================\n    case \"md_indicator\":\n      tokenType = \"macro\";\n      break;\n\n    case \"md_heading_text\":\n      tokenType = \"keyword\";\n      break;\n\n    default:\n      return null;\n  }\n\n  if (!tokenType) {\n    return null;\n  }\n\n  // For multiline tokens, only highlight the first line\n  const length =\n    startPosition.row === endPosition.row\n      ? endPosition.column - startPosition.column\n      : nodeText.indexOf(\"\\n\");\n\n  // Apply sourceMap to convert block-relative positions to file-absolute\n  // For the first line of the block, add both line and column offset\n  // For subsequent lines, only add line offset\n  const isFirstBlockLine = startPosition.row === 0;\n  const fileLine = sourceMap.lineOffset + startPosition.row;\n  const fileChar = isFirstBlockLine\n    ? sourceMap.columnOffset + startPosition.column\n    : startPosition.column;\n\n  return {\n    line: fileLine,\n    startChar: fileChar,\n    length: length > 0 ? length : nodeText.length,\n    tokenType: getTokenTypeIndex(tokenType),\n    tokenModifiers: getTokenModifiersMask(modifiers),\n  };\n}\n\n/**\n * Encode semantic tokens into the LSP delta format\n *\n * LSP expects tokens in a flattened array where each token is:\n * [deltaLine, deltaStartChar, length, tokenType, tokenModifiers]\n */\nexport function encodeSemanticTokens(tokens: SemanticToken[]): number[] {\n  const data: number[] = [];\n  let prevLine = 0;\n  let prevChar = 0;\n\n  for (const token of tokens) {\n    const deltaLine = token.line - prevLine;\n    const deltaChar = deltaLine === 0 ? token.startChar - prevChar : token.startChar;\n\n    data.push(deltaLine, deltaChar, token.length, token.tokenType, token.tokenModifiers);\n\n    prevLine = token.line;\n    prevChar = token.startChar;\n  }\n\n  return data;\n}\n"],"mappings":";;;;;;AASA,MAAa,aAAa;CACxB;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACD;;;;AAOD,MAAa,iBAAiB;CAC5B;CACA;CACA;CACA;CACD;;;;AAuBD,SAAgB,kBAAkB,MAAyB;AACzD,QAAO,WAAW,QAAQ,KAAK;;;;;AAMjC,SAAgB,sBAAsB,WAAoC;CACxE,IAAI,OAAO;AACX,MAAK,MAAM,OAAO,WAAW;EAC3B,MAAM,QAAQ,eAAe,QAAQ,IAAI;AACzC,MAAI,SAAS,EACX,SAAQ,KAAK;;AAGjB,QAAO;;;;;;AAOT,SAAgB,sBAAsB,UAA2C;CAC/E,MAAMA,SAA0B,EAAE;AAElC,MAAK,MAAM,SAAS,SAAS,OAC3B,uBAAsB,MAAM,MAAM,MAAM,WAAW,OAAO;AAI5D,QAAO,MAAM,GAAG,MAAM;AACpB,MAAI,EAAE,SAAS,EAAE,KACf,QAAO,EAAE,OAAO,EAAE;AAEpB,SAAO,EAAE,YAAY,EAAE;GACvB;AAEF,QAAO;;;;;AAMT,SAAS,sBACP,MACA,WACA,QACM;CAEN,MAAM,SAAU,KAAa,MAAM;CAGnC,MAAM,aAAa,SAA2B;EAC5C,MAAM,QAAQ,gBAAgB,MAAM,UAAU;AAC9C,MAAI,MACF,QAAO,KAAK,MAAM;AAIpB,MAAI,OAAO,gBAAgB,EAAE;AAC3B;AACE,cAAU,OAAO,YAAY;UACtB,OAAO,iBAAiB;AACjC,UAAO,YAAY;;;AAIvB,WAAU,OAAO,YAAY;;;;;;AAO/B,SAAS,gBAAgB,MAAkB,WAA4C;CACrF,IAAI,EAAE,MAAM,eAAe,gBAAgB;CAC3C,IAAI,WAAW,KAAK;AAGpB,KAAI,SAAS,QACX,QAAO;CAGT,IAAIC,YAA8B;CAClC,IAAIC,YAA6B,EAAE;AAEnC,SAAQ,MAAR;EAIE,KAAK;AACH,eAAY;AACZ;EAEF,KAAK;EACL,KAAK;EACL,KAAK;EACL,KAAK;AACH,eAAY;AACZ;EAEF,KAAK;EACL,KAAK;AACH,eAAY;AACZ;EAEF,KAAK;AAEH,eAAY;AACZ,eAAY,CAAC,aAAa;AAC1B;EAEF,KAAK;AACH,eAAY;AACZ;EAEF,KAAK;AACH,eAAY;AAEZ,OAAI,KAAK,QAAQ,SAAS,gBAAgB,KAAK,QAAQ,SAAS,eAC9D,aAAY,CAAC,cAAc;AAE7B;EAEF,KAAK;AACH,eAAY;AACZ;EAKF,KAAK;AACH,eAAY;AACZ;EAEF,KAAK;AACH,eAAY;AACZ;EAEF,KAAK;EACL,KAAK;AACH,eAAY;AACZ;EAKF,KAAK;EACL,KAAK,gBAAgB;AACnB,eAAY;GAEZ,MAAM,UAAU,SAAS,WAAW;AAEpC,OADmB,SAAS,SAAS,QAAQ,SAC5B,GAAG;AAElB,oBAAgB;KAAE,KAAK,YAAY;KAAK,QAAQ,YAAY,SAAS,QAAQ;KAAQ;AACrF,eAAW;;AAEb;;EAGF,KAAK;AACH,eAAY;AACZ,eAAY,CAAC,WAAW;AACxB;EAEF,KAAK;AACH,eAAY;AACZ;EAEF,KAAK;AACH,eAAY;AACZ;EAEF,KAAK;AACH,eAAY;AACZ,eAAY,CAAC,gBAAgB;AAC7B;EAKF,KAAK;EACL,KAAK;EACL,KAAK;EACL,KAAK;EACL,KAAK;EACL,KAAK;EACL,KAAK;EACL,KAAK;AACH,eAAY;AACZ;EAKF,KAAK;EACL,KAAK;AACH,eAAY;AACZ;EAEF,KAAK;AACH,eAAY;AACZ;EAKF,KAAK;AACH,eAAY;AACZ;EAKF,KAAK;AACH,eAAY;AACZ;EAEF,KAAK;AACH,eAAY;AACZ;EAEF,QACE,QAAO;;AAGX,KAAI,CAAC,UACH,QAAO;CAIT,MAAM,SACJ,cAAc,QAAQ,YAAY,MAC9B,YAAY,SAAS,cAAc,SACnC,SAAS,QAAQ,KAAK;CAK5B,MAAM,mBAAmB,cAAc,QAAQ;AAM/C,QAAO;EACL,MANe,UAAU,aAAa,cAAc;EAOpD,WANe,mBACb,UAAU,eAAe,cAAc,SACvC,cAAc;EAKhB,QAAQ,SAAS,IAAI,SAAS,SAAS;EACvC,WAAW,kBAAkB,UAAU;EACvC,gBAAgB,sBAAsB,UAAU;EACjD;;;;;;;;AASH,SAAgB,qBAAqB,QAAmC;CACtE,MAAMC,OAAiB,EAAE;CACzB,IAAI,WAAW;CACf,IAAI,WAAW;AAEf,MAAK,MAAM,SAAS,QAAQ;EAC1B,MAAM,YAAY,MAAM,OAAO;EAC/B,MAAM,YAAY,cAAc,IAAI,MAAM,YAAY,WAAW,MAAM;AAEvE,OAAK,KAAK,WAAW,WAAW,MAAM,QAAQ,MAAM,WAAW,MAAM,eAAe;AAEpF,aAAW,MAAM;AACjB,aAAW,MAAM;;AAGnB,QAAO"}

package/dist/services/synthesis.d.ts

@@ -0,0 +1,90 @@
+import { ActualizeEntry, InstanceEntry, SynthesisEntry } from "../ast/ast-types.js";
+import { Workspace } from "../model/workspace.js";
+import { Query, astQueryToModelQuery } from "./query.js";
+import { formatTimestamp } from "../formatters.js";
+
+//#region src/services/synthesis.d.ts
+/**
+ * Information about a synthesis definition including file context
+ */
+interface SynthesisInfo {
+  /** The synthesis entry */
+  entry: SynthesisEntry;
+  /** File containing this synthesis */
+  file: string;
+  /** The link ID for this synthesis */
+  linkId: string;
+  /** Title of the synthesis */
+  title: string;
+  /** Source queries for this synthesis */
+  sources: Query[];
+  /** The prompt text (if defined) */
+  prompt: string | null;
+}
+/**
+ * Information about an actualize entry including file context
+ */
+interface ActualizeInfo {
+  /** The actualize entry */
+  entry: ActualizeEntry;
+  /** File containing this actualize */
+  file: string;
+  /** Target synthesis link ID */
+  target: string;
+  /** Formatted timestamp of the actualize entry */
+  timestamp: string;
+}
+/**
+ * Extract source queries from a synthesis entry's metadata.
+ * Sources can be a single query or an array of queries.
+ */
+declare function getSynthesisSources(synthesis: SynthesisEntry): Query[];
+/**
+ * Extract the prompt text from a synthesis entry's content.
+ * Looks for content under a "# Prompt" header.
+ */
+declare function getSynthesisPrompt(synthesis: SynthesisEntry): string | null;
+/**
+ * Find all synthesis definitions in a workspace.
+ *
+ * @param workspace - The workspace to search
+ * @returns Array of synthesis info objects with file context
+ */
+declare function findAllSyntheses(workspace: Workspace): SynthesisInfo[];
+/**
+ * Find the latest actualize entry for a given synthesis by link ID.
+ * Searches across all files in the workspace.
+ *
+ * @param workspace - The workspace to search
+ * @param synthesisLinkId - The link ID of the synthesis to find actualizes for
+ * @returns The latest actualize info, or null if none found
+ */
+declare function findLatestActualize(workspace: Workspace, synthesisLinkId: string): ActualizeInfo | null;
+/**
+ * Get the 'updated' timestamp from an actualize entry's metadata.
+ *
+ * @param actualize - The actualize info (or null)
+ * @returns The updated timestamp string, or null if not found
+ */
+declare function getActualizeUpdatedTimestamp(actualize: ActualizeInfo | null): string | null;
+/**
+ * Find which file and model contain a specific entry.
+ * Uses object identity first, then falls back to location matching.
+ *
+ * @param workspace - The workspace to search
+ * @param entry - The entry to find
+ * @returns The file path, or undefined if not found
+ */
+declare function findEntryFile(workspace: Workspace, entry: InstanceEntry): string | undefined;
+/**
+ * Get the raw source text for an entry from its file.
+ * Requires providing a function to read file contents (to avoid filesystem dependency).
+ *
+ * @param entry - The entry to get text for
+ * @param source - The source text of the file containing the entry
+ * @returns The raw entry text
+ */
+declare function getEntrySourceText(entry: InstanceEntry, source: string): string;
+//#endregion
+export { ActualizeInfo, SynthesisInfo, findAllSyntheses, findEntryFile, findLatestActualize, getActualizeUpdatedTimestamp, getEntrySourceText, getSynthesisPrompt, getSynthesisSources };
+//# sourceMappingURL=synthesis.d.ts.map

package/dist/services/synthesis.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"synthesis.d.ts","names":[],"sources":["../../src/services/synthesis.ts"],"sourcesContent":[],"mappings":";;;;;;;;;AAciB,UAAA,aAAA,CAAa;EAkBb;EAqBD,KAAA,EArCP,cAqCO;EAsBA;EA8BA,IAAA,EAAA,MAAA;EA6BA;EAoCA,MAAA,EAAA,MAAA;EAgBA;EAwCA,KAAA,EAAA,MAAA;;WA1ML;;;;;;;UAQM,aAAA;;SAER;;;;;;;;;;;;iBAmBO,mBAAA,YAA+B,iBAAiB;;;;;iBAsBhD,kBAAA,YAA8B;;;;;;;iBA8B9B,gBAAA,YAA4B,YAAY;;;;;;;;;iBA6BxC,mBAAA,YACH,qCAEV;;;;;;;iBAiCa,4BAAA,YAAwC;;;;;;;;;iBAgBxC,aAAA,YAAyB,kBAAkB;;;;;;;;;iBAwC3C,kBAAA,QAA0B"}

package/dist/services/synthesis.js

@@ -0,0 +1,113 @@
+import { astQueryToModelQuery } from "./query.js";
+import { formatTimestamp } from "../formatters.js";
+
+//#region src/services/synthesis.ts
+/**
+* Extract source queries from a synthesis entry's metadata.
+* Sources can be a single query or an array of queries.
+*/
+function getSynthesisSources(synthesis) {
+	const sourcesMeta = synthesis.metadata.find((m) => m.key.value === "sources");
+	if (!sourcesMeta) return [];
+	const content = sourcesMeta.value.content;
+	if (content.type === "query_value") return [astQueryToModelQuery(content.query)];
+	if (content.type === "value_array") return content.elements.filter((e) => e.type === "query").map(astQueryToModelQuery);
+	return [];
+}
+/**
+* Extract the prompt text from a synthesis entry's content.
+* Looks for content under a "# Prompt" header.
+*/
+function getSynthesisPrompt(synthesis) {
+	if (!synthesis.content) return null;
+	let inPrompt = false;
+	const promptLines = [];
+	for (const child of synthesis.content.children) if (child.type === "markdown_header") if (child.text.toLowerCase().includes("prompt")) inPrompt = true;
+	else inPrompt = false;
+	else if (child.type === "content_line" && inPrompt) promptLines.push(child.text);
+	return promptLines.length > 0 ? promptLines.join("\n").trim() : null;
+}
+/**
+* Find all synthesis definitions in a workspace.
+*
+* @param workspace - The workspace to search
+* @returns Array of synthesis info objects with file context
+*/
+function findAllSyntheses(workspace) {
+	const syntheses = [];
+	for (const model of workspace.allModels()) for (const entry of model.ast.entries) if (entry.type === "synthesis_entry") syntheses.push({
+		entry,
+		file: model.file,
+		linkId: entry.header.linkId.id,
+		title: entry.header.title?.value ?? "(no title)",
+		sources: getSynthesisSources(entry),
+		prompt: getSynthesisPrompt(entry)
+	});
+	return syntheses;
+}
+/**
+* Find the latest actualize entry for a given synthesis by link ID.
+* Searches across all files in the workspace.
+*
+* @param workspace - The workspace to search
+* @param synthesisLinkId - The link ID of the synthesis to find actualizes for
+* @returns The latest actualize info, or null if none found
+*/
+function findLatestActualize(workspace, synthesisLinkId) {
+	let latest = null;
+	for (const model of workspace.allModels()) for (const entry of model.ast.entries) {
+		if (entry.type !== "actualize_entry") continue;
+		if (entry.header.target.id !== synthesisLinkId) continue;
+		const ts = formatTimestamp(entry.header.timestamp);
+		if (!latest || ts > latest.timestamp) latest = {
+			entry,
+			file: model.file,
+			target: entry.header.target.id,
+			timestamp: ts
+		};
+	}
+	return latest;
+}
+/**
+* Get the 'updated' timestamp from an actualize entry's metadata.
+*
+* @param actualize - The actualize info (or null)
+* @returns The updated timestamp string, or null if not found
+*/
+function getActualizeUpdatedTimestamp(actualize) {
+	if (!actualize) return null;
+	return actualize.entry.metadata.find((m) => m.key.value === "updated")?.value.raw ?? null;
+}
+/**
+* Find which file and model contain a specific entry.
+* Uses object identity first, then falls back to location matching.
+*
+* @param workspace - The workspace to search
+* @param entry - The entry to find
+* @returns The file path, or undefined if not found
+*/
+function findEntryFile(workspace, entry) {
+	for (const model of workspace.allModels()) for (const e of model.ast.entries) if (e === entry) return model.file;
+	const entryTs = formatTimestamp(entry.header.timestamp);
+	for (const model of workspace.allModels()) for (const e of model.ast.entries) {
+		if (e.type !== "instance_entry") continue;
+		if (formatTimestamp(e.header.timestamp) === entryTs && e.location.startIndex === entry.location.startIndex && e.location.endIndex === entry.location.endIndex) return model.file;
+	}
+}
+/**
+* Get the raw source text for an entry from its file.
+* Requires providing a function to read file contents (to avoid filesystem dependency).
+*
+* @param entry - The entry to get text for
+* @param source - The source text of the file containing the entry
+* @returns The raw entry text
+*/
+function getEntrySourceText(entry, source) {
+	const start = entry.location.startIndex;
+	const end = entry.location.endIndex;
+	return source.slice(start, end).trim();
+}
+
+//#endregion
+export { findAllSyntheses, findEntryFile, findLatestActualize, getActualizeUpdatedTimestamp, getEntrySourceText, getSynthesisPrompt, getSynthesisSources };
+//# sourceMappingURL=synthesis.js.map

package/dist/services/synthesis.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"synthesis.js","names":["promptLines: string[]","syntheses: SynthesisInfo[]","latest: ActualizeInfo | null"],"sources":["../../src/services/synthesis.ts"],"sourcesContent":["import type { Workspace } from \"../model/workspace.js\";\nimport type {\n  SynthesisEntry,\n  ActualizeEntry,\n  InstanceEntry,\n  Query as AstQuery,\n} from \"../ast/ast-types.js\";\nimport type { Query } from \"./query.js\";\nimport { formatTimestamp } from \"../formatters.js\";\nimport { astQueryToModelQuery } from \"./query.js\";\n\n/**\n * Information about a synthesis definition including file context\n */\nexport interface SynthesisInfo {\n  /** The synthesis entry */\n  entry: SynthesisEntry;\n  /** File containing this synthesis */\n  file: string;\n  /** The link ID for this synthesis */\n  linkId: string;\n  /** Title of the synthesis */\n  title: string;\n  /** Source queries for this synthesis */\n  sources: Query[];\n  /** The prompt text (if defined) */\n  prompt: string | null;\n}\n\n/**\n * Information about an actualize entry including file context\n */\nexport interface ActualizeInfo {\n  /** The actualize entry */\n  entry: ActualizeEntry;\n  /** File containing this actualize */\n  file: string;\n  /** Target synthesis link ID */\n  target: string;\n  /** Formatted timestamp of the actualize entry */\n  timestamp: string;\n}\n\n// Re-export formatTimestamp from formatters for backward compatibility\nexport { formatTimestamp } from \"../formatters.js\";\n\n// Re-export astQueryToModelQuery for backward compatibility\nexport { astQueryToModelQuery } from \"./query.js\";\n\n/**\n * Extract source queries from a synthesis entry's metadata.\n * Sources can be a single query or an array of queries.\n */\nexport function getSynthesisSources(synthesis: SynthesisEntry): Query[] {\n  const sourcesMeta = synthesis.metadata.find((m) => m.key.value === \"sources\");\n  if (!sourcesMeta) {\n    return [];\n  }\n\n  const content = sourcesMeta.value.content;\n  if (content.type === \"query_value\") {\n    return [astQueryToModelQuery(content.query)];\n  }\n  if (content.type === \"value_array\") {\n    return content.elements\n      .filter((e): e is AstQuery => e.type === \"query\")\n      .map(astQueryToModelQuery);\n  }\n  return [];\n}\n\n/**\n * Extract the prompt text from a synthesis entry's content.\n * Looks for content under a \"# Prompt\" header.\n */\nexport function getSynthesisPrompt(synthesis: SynthesisEntry): string | null {\n  if (!synthesis.content) {\n    return null;\n  }\n\n  let inPrompt = false;\n  const promptLines: string[] = [];\n\n  for (const child of synthesis.content.children) {\n    if (child.type === \"markdown_header\") {\n      const headerText = child.text.toLowerCase();\n      if (headerText.includes(\"prompt\")) {\n        inPrompt = true;\n      } else {\n        inPrompt = false;\n      }\n    } else if (child.type === \"content_line\" && inPrompt) {\n      promptLines.push(child.text);\n    }\n  }\n\n  return promptLines.length > 0 ? promptLines.join(\"\\n\").trim() : null;\n}\n\n/**\n * Find all synthesis definitions in a workspace.\n *\n * @param workspace - The workspace to search\n * @returns Array of synthesis info objects with file context\n */\nexport function findAllSyntheses(workspace: Workspace): SynthesisInfo[] {\n  const syntheses: SynthesisInfo[] = [];\n\n  for (const model of workspace.allModels()) {\n    for (const entry of model.ast.entries) {\n      if (entry.type === \"synthesis_entry\") {\n        syntheses.push({\n          entry,\n          file: model.file,\n          linkId: entry.header.linkId.id,\n          title: entry.header.title?.value ?? \"(no title)\",\n          sources: getSynthesisSources(entry),\n          prompt: getSynthesisPrompt(entry),\n        });\n      }\n    }\n  }\n\n  return syntheses;\n}\n\n/**\n * Find the latest actualize entry for a given synthesis by link ID.\n * Searches across all files in the workspace.\n *\n * @param workspace - The workspace to search\n * @param synthesisLinkId - The link ID of the synthesis to find actualizes for\n * @returns The latest actualize info, or null if none found\n */\nexport function findLatestActualize(\n  workspace: Workspace,\n  synthesisLinkId: string,\n): ActualizeInfo | null {\n  let latest: ActualizeInfo | null = null;\n\n  for (const model of workspace.allModels()) {\n    for (const entry of model.ast.entries) {\n      if (entry.type !== \"actualize_entry\") {\n        continue;\n      }\n      if (entry.header.target.id !== synthesisLinkId) {\n        continue;\n      }\n\n      const ts = formatTimestamp(entry.header.timestamp);\n      if (!latest || ts > latest.timestamp) {\n        latest = {\n          entry,\n          file: model.file,\n          target: entry.header.target.id,\n          timestamp: ts,\n        };\n      }\n    }\n  }\n\n  return latest;\n}\n\n/**\n * Get the 'updated' timestamp from an actualize entry's metadata.\n *\n * @param actualize - The actualize info (or null)\n * @returns The updated timestamp string, or null if not found\n */\nexport function getActualizeUpdatedTimestamp(actualize: ActualizeInfo | null): string | null {\n  if (!actualize) {\n    return null;\n  }\n  const updated = actualize.entry.metadata.find((m) => m.key.value === \"updated\");\n  return updated?.value.raw ?? null;\n}\n\n/**\n * Find which file and model contain a specific entry.\n * Uses object identity first, then falls back to location matching.\n *\n * @param workspace - The workspace to search\n * @param entry - The entry to find\n * @returns The file path, or undefined if not found\n */\nexport function findEntryFile(workspace: Workspace, entry: InstanceEntry): string | undefined {\n  // First try object identity (fast path when entry came from this workspace)\n  for (const model of workspace.allModels()) {\n    for (const e of model.ast.entries) {\n      if (e === entry) {\n        return model.file;\n      }\n    }\n  }\n\n  // Fallback: match by location and timestamp (for entries that may have been cloned)\n  const entryTs = formatTimestamp(entry.header.timestamp);\n  for (const model of workspace.allModels()) {\n    for (const e of model.ast.entries) {\n      if (e.type !== \"instance_entry\") {\n        continue;\n      }\n      // Match by timestamp and location for uniqueness\n      const eTs = formatTimestamp(e.header.timestamp);\n      if (\n        eTs === entryTs &&\n        e.location.startIndex === entry.location.startIndex &&\n        e.location.endIndex === entry.location.endIndex\n      ) {\n        return model.file;\n      }\n    }\n  }\n\n  return undefined;\n}\n\n/**\n * Get the raw source text for an entry from its file.\n * Requires providing a function to read file contents (to avoid filesystem dependency).\n *\n * @param entry - The entry to get text for\n * @param source - The source text of the file containing the entry\n * @returns The raw entry text\n */\nexport function getEntrySourceText(entry: InstanceEntry, source: string): string {\n  const start = entry.location.startIndex;\n  const end = entry.location.endIndex;\n  return source.slice(start, end).trim();\n}\n"],"mappings":";;;;;;;;AAqDA,SAAgB,oBAAoB,WAAoC;CACtE,MAAM,cAAc,UAAU,SAAS,MAAM,MAAM,EAAE,IAAI,UAAU,UAAU;AAC7E,KAAI,CAAC,YACH,QAAO,EAAE;CAGX,MAAM,UAAU,YAAY,MAAM;AAClC,KAAI,QAAQ,SAAS,cACnB,QAAO,CAAC,qBAAqB,QAAQ,MAAM,CAAC;AAE9C,KAAI,QAAQ,SAAS,cACnB,QAAO,QAAQ,SACZ,QAAQ,MAAqB,EAAE,SAAS,QAAQ,CAChD,IAAI,qBAAqB;AAE9B,QAAO,EAAE;;;;;;AAOX,SAAgB,mBAAmB,WAA0C;AAC3E,KAAI,CAAC,UAAU,QACb,QAAO;CAGT,IAAI,WAAW;CACf,MAAMA,cAAwB,EAAE;AAEhC,MAAK,MAAM,SAAS,UAAU,QAAQ,SACpC,KAAI,MAAM,SAAS,kBAEjB,KADmB,MAAM,KAAK,aAAa,CAC5B,SAAS,SAAS,CAC/B,YAAW;KAEX,YAAW;UAEJ,MAAM,SAAS,kBAAkB,SAC1C,aAAY,KAAK,MAAM,KAAK;AAIhC,QAAO,YAAY,SAAS,IAAI,YAAY,KAAK,KAAK,CAAC,MAAM,GAAG;;;;;;;;AASlE,SAAgB,iBAAiB,WAAuC;CACtE,MAAMC,YAA6B,EAAE;AAErC,MAAK,MAAM,SAAS,UAAU,WAAW,CACvC,MAAK,MAAM,SAAS,MAAM,IAAI,QAC5B,KAAI,MAAM,SAAS,kBACjB,WAAU,KAAK;EACb;EACA,MAAM,MAAM;EACZ,QAAQ,MAAM,OAAO,OAAO;EAC5B,OAAO,MAAM,OAAO,OAAO,SAAS;EACpC,SAAS,oBAAoB,MAAM;EACnC,QAAQ,mBAAmB,MAAM;EAClC,CAAC;AAKR,QAAO;;;;;;;;;;AAWT,SAAgB,oBACd,WACA,iBACsB;CACtB,IAAIC,SAA+B;AAEnC,MAAK,MAAM,SAAS,UAAU,WAAW,CACvC,MAAK,MAAM,SAAS,MAAM,IAAI,SAAS;AACrC,MAAI,MAAM,SAAS,kBACjB;AAEF,MAAI,MAAM,OAAO,OAAO,OAAO,gBAC7B;EAGF,MAAM,KAAK,gBAAgB,MAAM,OAAO,UAAU;AAClD,MAAI,CAAC,UAAU,KAAK,OAAO,UACzB,UAAS;GACP;GACA,MAAM,MAAM;GACZ,QAAQ,MAAM,OAAO,OAAO;GAC5B,WAAW;GACZ;;AAKP,QAAO;;;;;;;;AAST,SAAgB,6BAA6B,WAAgD;AAC3F,KAAI,CAAC,UACH,QAAO;AAGT,QADgB,UAAU,MAAM,SAAS,MAAM,MAAM,EAAE,IAAI,UAAU,UAAU,EAC/D,MAAM,OAAO;;;;;;;;;;AAW/B,SAAgB,cAAc,WAAsB,OAA0C;AAE5F,MAAK,MAAM,SAAS,UAAU,WAAW,CACvC,MAAK,MAAM,KAAK,MAAM,IAAI,QACxB,KAAI,MAAM,MACR,QAAO,MAAM;CAMnB,MAAM,UAAU,gBAAgB,MAAM,OAAO,UAAU;AACvD,MAAK,MAAM,SAAS,UAAU,WAAW,CACvC,MAAK,MAAM,KAAK,MAAM,IAAI,SAAS;AACjC,MAAI,EAAE,SAAS,iBACb;AAIF,MADY,gBAAgB,EAAE,OAAO,UAAU,KAErC,WACR,EAAE,SAAS,eAAe,MAAM,SAAS,cACzC,EAAE,SAAS,aAAa,MAAM,SAAS,SAEvC,QAAO,MAAM;;;;;;;;;;;AAgBrB,SAAgB,mBAAmB,OAAsB,QAAwB;CAC/E,MAAM,QAAQ,MAAM,SAAS;CAC7B,MAAM,MAAM,MAAM,SAAS;AAC3B,QAAO,OAAO,MAAM,OAAO,IAAI,CAAC,MAAM"}

package/dist/source-map.d.ts

@@ -0,0 +1,42 @@
+//#region src/source-map.d.ts
+/**
+ * A simple position (line and column).
+ * Both are 0-based to match tree-sitter's Point type.
+ */
+interface Position {
+  /** 0-based line number */
+  line: number;
+  /** 0-based column number */
+  column: number;
+}
+/**
+ * SourceMap tracks the relationship between block-relative positions
+ * (as returned by tree-sitter) and file-absolute positions.
+ *
+ * For standalone .thalo files, this is an identity map (all offsets are 0).
+ * For embedded blocks in markdown, this contains the offset to the block start.
+ */
+interface SourceMap {
+  /** Character offset where block content starts in the file */
+  readonly charOffset: number;
+  /** Line number where block content starts (0-based) */
+  readonly lineOffset: number;
+  /** Column offset on the starting line (usually 0 for fenced code blocks) */
+  readonly columnOffset: number;
+  /** Number of lines in the block content */
+  readonly lineCount: number;
+}
+/**
+ * Result of finding a block at a position.
+ */
+interface BlockMatch<T extends {
+  sourceMap: SourceMap;
+}> {
+  /** The matched block */
+  block: T;
+  /** Position relative to the block start (0-based line and column) */
+  blockPosition: Position;
+}
+//#endregion
+export { BlockMatch, Position, SourceMap };
+//# sourceMappingURL=source-map.d.ts.map

package/dist/source-map.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"source-map.d.ts","names":[],"sources":["../src/source-map.ts"],"sourcesContent":[],"mappings":";AAMA;AAuBA;AA8OA;;AAES,UAvQQ,QAAA,CAuQR;EAEQ;EAAQ,IAAA,EAAA,MAAA;;;;;;;;;;;UAlPR,SAAA;;;;;;;;;;;;;UA8OA;aAAkC;;;SAE1C;;iBAEQ"}