@glion/parser 0.16.0 → 0.20.0

package/README.md

@@ -60,6 +60,16 @@ const customTree = unified()
 
 Options accept a partial `Delimiters` object, so only the characters you want to override need to be provided.
 
+### `Hl7v2Processor`
+
+The `unified` `Processor<Root, Root | undefined, Root | undefined>` type every HL7v2 pipeline satisfies — head and tail admit `undefined` so the bare `unified().use(hl7v2Parser).freeze()` is assignable. Use it to type a processor built with `unified().use(hl7v2Parser)` regardless of which transformer or compiler plugins are added. `@glion/mllp` re-exports it unchanged.
+
+```ts
+import type { Hl7v2Processor } from "@glion/parser";
+
+const processor: Hl7v2Processor = unified().use(hl7v2Parser).freeze();
+```
+
 ## Parsing model
 
 - **Pull-based tokenizer.** Single pass, minimal object allocations — suitable for high-throughput ingestion.

package/dist/index.js.map

@@ -1 +1 @@
-{"version":3,"file":"index.js","names":[],"sources":["../src/preprocessor.ts","../src/processor.ts","../src/tokenizer.ts","../src/parser.ts"],"sourcesContent":["import { DEFAULT_DELIMITERS } from \"@glion/utils\";\n\nimport type { ParserContext, PreprocessorStep } from \"./types\";\n\n// PreprocessorStep is declared in `types.ts` to avoid circular imports\n\n/**\n * Step: strip UTF-8 BOM.\n */\nexport const stripBOM: PreprocessorStep = (ctx) => {\n  if (ctx.input.codePointAt(0) === 0xfe_ff) {\n    ctx.input = ctx.input.slice(1);\n  }\n  return ctx;\n};\n\n/**\n * Step: normalize newlines to the default delimiters.\n */\nexport const normalizeNewlines: PreprocessorStep = (ctx) => {\n  ctx.input = ctx.input.replaceAll(/\\r?\\n/g, ctx.delimiters.segment);\n  return ctx;\n};\n\n/**\n * Step: auto-detect delimiters from MSH.\n */\nexport const detectDelimiters: PreprocessorStep = (ctx) => {\n  if (ctx.input.startsWith(\"MSH\")) {\n    const field = ctx.input.charAt(3) || DEFAULT_DELIMITERS.field;\n    const enc = ctx.input.slice(4, 8);\n    const component = enc.charAt(0) || DEFAULT_DELIMITERS.component;\n    const repetition = enc.charAt(1) || DEFAULT_DELIMITERS.repetition;\n    const _escape = enc.charAt(2) || DEFAULT_DELIMITERS.escape;\n    const subcomponent = enc.charAt(3) || DEFAULT_DELIMITERS.subcomponent;\n    const { segment } = DEFAULT_DELIMITERS;\n\n    ctx.delimiters = {\n      component,\n      escape: _escape,\n      field,\n      repetition,\n      segment,\n      subcomponent,\n    };\n  }\n  return ctx;\n};\n\n/**\n * Default preprocessing pipeline.\n */\nexport const defaultPreprocessors: PreprocessorStep[] = [\n  stripBOM,\n  normalizeNewlines,\n  detectDelimiters,\n];\n\n/**\n * Run the pipeline to initialize ParserContext.\n */\nexport function runPreprocessors(\n  ctx: ParserContext,\n  steps: PreprocessorStep[]\n): ParserContext {\n  for (const step of steps) {\n    // oxlint-disable-next-line no-param-reassign\n    ctx = step(ctx);\n  }\n\n  return ctx;\n}\n","import type {\n  Component,\n  Field,\n  FieldRepetition,\n  Root,\n  Segment,\n  Subcomponent,\n} from \"@glion/ast\";\nimport { isEmptyNode } from \"@glion/utils\";\nimport type { Position } from \"unist\";\n\nimport type { Token } from \"./types\";\n\n/**\n * Check if a field has structural content beyond a single empty value.\n * A field with multiple repetitions, multiple components, or multiple\n * subcomponents has structural content even if all values are empty.\n * This is used to prevent dropping trailing fields like |^| or |&|\n * that have meaningful delimiter structure.\n */\nfunction hasStructuralContent(field: Field): boolean {\n  if (field.children.length > 1) {\n    return true;\n  }\n  const rep = field.children[0];\n  if (rep && rep.children.length > 1) {\n    return true;\n  }\n  const comp = rep?.children[0];\n  return comp !== undefined && comp.children.length > 1;\n}\n\nfunction createSubcomponent(start: Position[\"start\"]): Subcomponent {\n  return {\n    position: { end: start, start },\n    type: \"subcomponent\",\n    value: \"\",\n  };\n}\n\nfunction createComponent(start: Position[\"start\"]): Component {\n  return {\n    children: [],\n    position: { end: start, start },\n    type: \"component\",\n  };\n}\n\nfunction createFieldRepetition(start: Position[\"start\"]): FieldRepetition {\n  return {\n    children: [],\n    position: { end: start, start },\n    type: \"field-repetition\",\n  };\n}\n\nfunction createField(start: Position[\"start\"]): Field {\n  return {\n    children: [],\n    position: { end: start, start },\n    type: \"field\",\n  };\n}\n\n// Shared core: process a single token into mutable parse state\nfunction createParserCore() {\n  const root: Root = {\n    children: [],\n    type: \"root\",\n  };\n\n  let seg: Segment | null = null;\n  let field: Field | null = null;\n  let rep: FieldRepetition | null = null;\n  let comp: Component | null = null;\n  let currentSub: Subcomponent | null = null;\n  let segmentHasContent = false;\n  let lastContentEnd: Position[\"end\"] | null = null;\n  let documentEnd: Position[\"end\"] | null = null; // Track document-level end position\n  let expectingSegmentName = true; // Start expecting a segment name\n  let justSetSegmentName = false; // Track if we just set the segment name\n  let rootStartSet = false; // Track if we've set the root start position\n\n  const resetState = () => {\n    seg = null;\n    field = null;\n    rep = null;\n    comp = null;\n    currentSub = null;\n    segmentHasContent = false;\n    lastContentEnd = null;\n    expectingSegmentName = true;\n  };\n\n  const openSegment = (name: string, position: Position) => {\n    seg = {\n      children: [],\n      name,\n      position: { end: position.end, start: position.start },\n      type: \"segment\",\n    };\n    root.children.push(seg);\n    field = null;\n    rep = null;\n    comp = null;\n    currentSub = null;\n    segmentHasContent = false;\n    justSetSegmentName = true;\n    expectingSegmentName = false;\n  };\n\n  /**\n   * Ensure there is an active segment. If delimiter tokens arrive before\n   * any segment name, lazily open a segment with an empty name.\n   */\n  const ensureSegment = (start: Position[\"start\"]) => {\n    if (!seg) {\n      openSegment(\"\", { end: start, start });\n      // Reset justSetSegmentName so the first FIELD_DELIM isn't skipped\n      // (there's no real segment name to separate from fields)\n      justSetSegmentName = false;\n    }\n  };\n\n  const openField = (start: Position[\"start\"]) => {\n    ensureSegment(start);\n    field = createField(start);\n    seg!.children.push(field);\n    rep = null;\n    comp = null;\n    currentSub = null;\n    segmentHasContent = true;\n  };\n\n  const openRepetition = (start: Position[\"start\"]) => {\n    if (!field) {\n      openField(start);\n      // Add an empty first repetition before the delimiter\n      const emptyRep = createFieldRepetition(start);\n      field!.children.push(emptyRep);\n    }\n    // If field has no children yet, add empty first rep\n    if (field!.children.length === 0) {\n      const emptyRep = createFieldRepetition(start);\n      field!.children.push(emptyRep);\n    }\n    rep = createFieldRepetition(start);\n    field!.children.push(rep);\n    comp = null;\n    currentSub = null;\n    segmentHasContent = true;\n  };\n\n  // Ensure there's a component to add subcomponents to (for SUBCOMP_DELIM)\n  // Does NOT add empty sibling components.\n  const ensureComponent = (start: Position[\"start\"]) => {\n    if (!field) {\n      openField(start);\n    }\n    if (!rep) {\n      rep = createFieldRepetition(start);\n      field!.children.push(rep);\n    }\n    if (!comp) {\n      comp = createComponent(start);\n      rep.children.push(comp);\n    }\n    segmentHasContent = true;\n  };\n\n  const openComponent = (start: Position[\"start\"]) => {\n    if (!field) {\n      openField(start);\n      // Build out a rep + empty first component before the delimiter\n      rep = createFieldRepetition(start);\n      field!.children.push(rep);\n      const emptyComp = createComponent(start);\n      rep.children.push(emptyComp);\n    }\n    if (!rep) {\n      rep = createFieldRepetition(start);\n      field!.children.push(rep);\n      const emptyComp = createComponent(start);\n      rep.children.push(emptyComp);\n    } else if (rep.children.length === 0) {\n      const emptyComp = createComponent(start);\n      rep.children.push(emptyComp);\n    }\n    comp = createComponent(start);\n    rep.children.push(comp);\n    currentSub = null;\n    segmentHasContent = true;\n  };\n\n  const ensureForText = (start: Position[\"start\"]) => {\n    if (!field) {\n      openField(start);\n    }\n    if (!rep) {\n      rep = createFieldRepetition(start);\n      field!.children.push(rep);\n    }\n    if (!comp) {\n      comp = createComponent(start);\n      rep.children.push(comp);\n    }\n    if (!currentSub) {\n      currentSub = createSubcomponent(start);\n      comp.children.push(currentSub);\n      segmentHasContent = true;\n    }\n  };\n\n  const updatePositionsToEnd = (endPos: Position[\"end\"]) => {\n    if (currentSub?.position) {\n      currentSub.position.end = endPos;\n    }\n    if (comp?.position) {\n      comp.position.end = endPos;\n    }\n    if (rep?.position) {\n      rep.position.end = endPos;\n    }\n    if (field?.position) {\n      field.position.end = endPos;\n    }\n    if (seg?.position) {\n      seg.position.end = endPos;\n    }\n  };\n\n  const processToken = (tok: Token) => {\n    switch (tok.type) {\n      case \"SEGMENT_END\": {\n        const endPos = lastContentEnd || tok.position.start;\n        updatePositionsToEnd(endPos);\n        dropTrailingEmptyFieldIfPresent();\n        resetState();\n        return;\n      }\n      case \"FIELD_DELIM\": {\n        lastContentEnd = tok.position.end;\n        documentEnd = tok.position.end;\n        // Skip the first field delimiter after setting segment name\n        if (justSetSegmentName) {\n          justSetSegmentName = false;\n          return;\n        }\n        // Leading field delimiter implies an empty first field\n        if (!field) {\n          openField(tok.position.start);\n        }\n        openField(tok.position.end);\n        return;\n      }\n      case \"REPETITION_DELIM\": {\n        lastContentEnd = tok.position.end;\n        documentEnd = tok.position.end;\n        if (!field) {\n          openField(tok.position.start);\n        }\n        openRepetition(tok.position.end);\n        return;\n      }\n      case \"COMPONENT_DELIM\": {\n        lastContentEnd = tok.position.end;\n        documentEnd = tok.position.end;\n        openComponent(tok.position.end);\n        return;\n      }\n      case \"SUBCOMP_DELIM\": {\n        lastContentEnd = tok.position.end;\n        documentEnd = tok.position.end;\n        ensureComponent(tok.position.start);\n        // If comp has no children yet, add empty first subcomponent\n        if (comp!.children.length === 0) {\n          const emptySub = createSubcomponent(tok.position.start);\n          comp!.children.push(emptySub);\n        }\n        currentSub = createSubcomponent(tok.position.end);\n        comp?.children.push(currentSub);\n        segmentHasContent = true;\n        return;\n      }\n      case \"TEXT\": {\n        const val = tok.value ?? \"\";\n\n        // Set root start position on first token\n        if (!rootStartSet) {\n          root.position = {\n            end: tok.position.start,\n            start: tok.position.start,\n          };\n          rootStartSet = true;\n        }\n\n        // If we're expecting a segment name, this TEXT is the segment name\n        if (expectingSegmentName) {\n          openSegment(val, tok.position);\n          lastContentEnd = tok.position.end;\n          documentEnd = tok.position.end;\n          return;\n        }\n\n        // Otherwise, it's regular field content\n        ensureForText(tok.position.start);\n        if (currentSub) {\n          currentSub.value += val;\n        }\n        updatePositionsToEnd(tok.position.end);\n        lastContentEnd = tok.position.end;\n        documentEnd = tok.position.end;\n        segmentHasContent = true;\n        return;\n      }\n      default: {\n        throw new Error(`Unexpected token type: ${tok.type}`);\n      }\n    }\n  };\n\n  // Do not pre-open a segment; lazily open upon first non-SEGMENT_END token.\n\n  const finalize = () => {\n    if (lastContentEnd && seg) {\n      updatePositionsToEnd(lastContentEnd);\n    }\n    dropTrailingEmptyFieldIfPresent();\n    if (seg && !segmentHasContent && !seg.name) {\n      // Drop trailing empty segment only when it was lazily opened by\n      // `ensureSegment` (no name, no content). A segment that arrived\n      // with a name carries information — even if no fields followed —\n      // and dropping it would break round-trip idempotency: the\n      // serializer re-emits the name, the next parse would drop it\n      // again, and any sibling segment immediately preceding it would\n      // shift into the trailing position and get dropped on the\n      // following parse.\n      root.children.pop();\n    }\n\n    // Set root end position once at finalization using document-level end\n    if (root.position && documentEnd) {\n      root.position.end = documentEnd;\n    } else if (!root.position) {\n      root.position = {\n        end: { column: 1, line: 1, offset: 0 },\n        start: { column: 1, line: 1, offset: 0 },\n      };\n    }\n\n    return root;\n  };\n\n  /**\n   * Drop only the phantom trailing empty field created by a terminal field\n   * delimiter (e.g. the last `|` in `PID|foo|`). Without this, every segment\n   * ending with `|` would gain an extra empty field on each parse cycle.\n   */\n  function dropTrailingEmptyFieldIfPresent() {\n    if (!seg || seg.children.length === 0) {\n      return;\n    }\n    const lastChild = seg.children.at(-1);\n    if (lastChild?.type !== \"field\") {\n      return;\n    }\n    const lastField = lastChild as Field;\n    // A field like `|^|` is isEmpty (no leaf data) but has structural content\n    // (component separators) and must be preserved.\n    if (isEmptyNode(lastField) && !hasStructuralContent(lastField)) {\n      seg.children.pop();\n    }\n  }\n\n  return { finalize, processToken, root };\n}\n\nexport function parseHL7v2FromIterator(tokens: Iterable<Token>): Root {\n  const core = createParserCore();\n  for (const tok of tokens) {\n    core.processToken(tok);\n  }\n  return core.finalize();\n}\n","// src/tokenizer.ts\nimport type { Delimiters } from \"@glion/ast\";\nimport type { Position } from \"unist\";\n\nimport type { ParserContext, Token, Tokenizer, TokenType } from \"./types\";\n\nconst MSH_SEGMENT_START = 0;\nconst MSH_SEGMENT_END = 3;\nconst MSH_FIELD_SEPERATOR_START = 3;\nconst MSH_FIELD_SEPERATOR_END = 4;\nconst MSH_FIELD_DELIMITER_START = 4;\nconst MSH_FIELD_DELIMITER_END = 8;\n\nexport class HL7v2Tokenizer implements Tokenizer, Iterable<Token> {\n  private input = \"\";\n  private i = 0;\n  private line = 1;\n  private col = 1;\n  private delims!: Delimiters;\n\n  // Only-run-once MSH bootstrap at the start of the file\n  private didMshBootstrap = false;\n  private pendingBootstrap: null | Array<\n    | { kind: \"TEXT\"; value: string; advance: number }\n    | { kind: \"FIELD_DELIM\"; advance: number }\n  > = null; // queue to emit TEXT('MSH'), FIELD_DELIM, TEXT('^~\\\\&')\n\n  reset(ctx: ParserContext) {\n    this.input = ctx.input;\n    this.delims = ctx.delimiters;\n    this.i = 0;\n    this.line = 1;\n    this.col = 1;\n    this.didMshBootstrap = false;\n    this.pendingBootstrap = null;\n\n    // Prepare a one-time bootstrap if file starts with MSH\n    if (this.input.startsWith(\"MSH\")) {\n      // Precompute MSH and MSH.2; MSH.1 is the field delimiter char at index 3\n      const msh = this._slice(MSH_SEGMENT_START, MSH_SEGMENT_END);\n      const msh1 = this._slice(\n        MSH_FIELD_SEPERATOR_START,\n        MSH_FIELD_SEPERATOR_END\n      ); // the field delimiter char at index 3\n      const msh2 = this._slice(\n        MSH_FIELD_DELIMITER_START,\n        MSH_FIELD_DELIMITER_END\n      ); // may be shorter than 4 if truncated\n      this.pendingBootstrap = [\n        { advance: msh.length, kind: \"TEXT\", value: msh },\n        { advance: 0, kind: \"FIELD_DELIM\" },\n        { advance: msh1.length, kind: \"TEXT\", value: msh1 },\n        { advance: 0, kind: \"FIELD_DELIM\" },\n        { advance: msh2.length, kind: \"TEXT\", value: msh2 },\n      ];\n      // NOTE: we have not advanced indices yet; we will as we emit tokens\n    }\n  }\n\n  // oxlint-disable-next-line complexity\n  next(): Token | null {\n    const s = this.input;\n    const n = s.length;\n    if (this.i >= n && !this.pendingBootstrap?.length) {\n      return null;\n    }\n\n    const start = { column: this.col, line: this.line, offset: this.i };\n\n    // ---- One-time MSH bootstrap at file start ----\n    if (!this.didMshBootstrap && this.pendingBootstrap) {\n      const step = this.pendingBootstrap.shift();\n      if (step) {\n        if (step.kind === \"TEXT\") {\n          const take = Math.min(step.advance, n - this.i);\n          const out = step.value.slice(0, take);\n          this._fastAdvance(out);\n          if (this.pendingBootstrap.length === 0) {\n            this.pendingBootstrap = null;\n            this.didMshBootstrap = true;\n          }\n          return this._tok(\"TEXT\", out, start);\n        }\n        // FIELD_DELIM: advance exactly one char (the delimiter) and emit a FIELD_DELIM token\n        this._advance(Math.min(step.advance, n - this.i));\n        if (this.pendingBootstrap.length === 0) {\n          this.pendingBootstrap = null;\n          this.didMshBootstrap = true;\n        }\n        return this._tok(\"FIELD_DELIM\", undefined, start);\n      }\n    }\n\n    // ---- Normal tokenization (delimiters + text) ----\n    if (this.i >= n) {\n      return null;\n    }\n    const ch = s.charAt(this.i);\n\n    if (ch === this.delims.segment) {\n      this._advance(1, true);\n      return this._tok(\"SEGMENT_END\", undefined, start);\n    }\n    if (ch === this.delims.field) {\n      this._advance(1);\n      return this._tok(\"FIELD_DELIM\", undefined, start);\n    }\n    if (ch === this.delims.repetition) {\n      this._advance(1);\n      return this._tok(\"REPETITION_DELIM\", undefined, start);\n    }\n    if (ch === this.delims.component) {\n      this._advance(1);\n      return this._tok(\"COMPONENT_DELIM\", undefined, start);\n    }\n    if (ch === this.delims.subcomponent) {\n      this._advance(1);\n      return this._tok(\"SUBCOMP_DELIM\", undefined, start);\n    }\n\n    // TEXT until next delimiter or end\n    let j = this.i;\n    const cmp = this.delims.component,\n      fld = this.delims.field,\n      rep = this.delims.repetition,\n      seg = this.delims.segment,\n      sub = this.delims.subcomponent;\n\n    while (j < s.length) {\n      const c = s.charAt(j);\n      if (c === seg || c === fld || c === rep || c === cmp || c === sub) {\n        break;\n      }\n      j += 1;\n    }\n\n    const val = s.slice(this.i, j);\n    this._fastAdvance(val);\n    return this._tok(\"TEXT\", val, start);\n  }\n\n  // ---- helpers ----\n  private _slice(start: number, end: number): string {\n    return this.input.slice(start, Math.min(end, this.input.length));\n  }\n\n  private _advance(n: number, isNewline = false) {\n    this.i += n;\n    if (isNewline) {\n      this.line += 1;\n      this.col = 1;\n    } else {\n      this.col += n;\n    }\n  }\n\n  private _fastAdvance(chunk: string) {\n    const parts = chunk.split(this.delims.segment);\n    if (parts.length > 1) {\n      this.line += parts.length - 1;\n      this.col = (parts.at(-1)?.length ?? 0) + 1;\n    } else {\n      this.col += chunk.length;\n    }\n    this.i += chunk.length;\n  }\n\n  private _tok(\n    type: TokenType,\n    value: string | undefined,\n    start: Position[\"start\"]\n  ): Token {\n    const end = { column: this.col, line: this.line, offset: this.i };\n    return { position: { end, start }, type, value };\n  }\n\n  // Iterable protocol (sync)\n  [Symbol.iterator](): Iterator<Token> {\n    return {\n      next: () => {\n        const t = this.next();\n        if (!t) {\n          return { done: true, value: undefined as unknown as Token };\n        }\n        return { done: false, value: t };\n      },\n    };\n  }\n}\n","import type { Root } from \"@glion/ast\";\nimport type { HL7v2Settings } from \"@glion/config\";\nimport { DEFAULT_DELIMITERS } from \"@glion/utils\";\nimport type { Plugin, Processor } from \"unified\";\n\nimport { defaultPreprocessors, runPreprocessors } from \"./preprocessor\";\nimport { parseHL7v2FromIterator } from \"./processor\";\nimport { HL7v2Tokenizer } from \"./tokenizer\";\nimport type { ParseOptions, ParserContext, Token, Tokenizer } from \"./types\";\n\nfunction* iterateTokenizerSync(t: Tokenizer): Iterable<Token> {\n  for (let tok = t.next(); tok; tok = t.next()) {\n    yield tok;\n  }\n}\n\nexport function parseHL7v2(\n  input: string,\n  opts: ParseOptions = {},\n  settings?: HL7v2Settings\n): Root {\n  // Merge settings: defaults < provided settings\n  // Note: MSH auto-detection (via detectDelimiters preprocessor) runs after this\n  // and will override delimiters if the message starts with MSH\n\n  let ctx: ParserContext = {\n    delimiters: {\n      ...DEFAULT_DELIMITERS,\n      ...(settings?.delimiters ?? {}),\n    },\n    input,\n  };\n  // Run preprocessing\n  ctx = runPreprocessors(ctx, opts.preprocess || defaultPreprocessors);\n  // Run tokenizer\n  const tokenizer = new HL7v2Tokenizer();\n  // Reset tokenizer.\n  tokenizer.reset(ctx);\n  // Parse\n  return parseHL7v2FromIterator(iterateTokenizerSync(tokenizer));\n}\n\nconst hl7v2Parser: Plugin<[ParseOptions?], string, Root> = function hl7v2Parser(\n  this: Processor,\n  options = {}\n) {\n  this.parser = (document: string) => {\n    // Get settings from processor (populated by unified-args from config)\n    const settings = this.data(\"settings\");\n\n    return parseHL7v2(document, options, settings);\n  };\n};\n\nexport default hl7v2Parser;\n"],"mappings":";;;;;AASA,MAAa,YAA8B,QAAQ;AACjD,KAAI,IAAI,MAAM,YAAY,EAAE,KAAK,MAC/B,KAAI,QAAQ,IAAI,MAAM,MAAM,EAAE;AAEhC,QAAO;;;;;AAMT,MAAa,qBAAuC,QAAQ;AAC1D,KAAI,QAAQ,IAAI,MAAM,WAAW,UAAU,IAAI,WAAW,QAAQ;AAClE,QAAO;;;;;AAMT,MAAa,oBAAsC,QAAQ;AACzD,KAAI,IAAI,MAAM,WAAW,MAAM,EAAE;EAC/B,MAAM,QAAQ,IAAI,MAAM,OAAO,EAAE,IAAI,mBAAmB;EACxD,MAAM,MAAM,IAAI,MAAM,MAAM,GAAG,EAAE;EACjC,MAAM,YAAY,IAAI,OAAO,EAAE,IAAI,mBAAmB;EACtD,MAAM,aAAa,IAAI,OAAO,EAAE,IAAI,mBAAmB;EACvD,MAAM,UAAU,IAAI,OAAO,EAAE,IAAI,mBAAmB;EACpD,MAAM,eAAe,IAAI,OAAO,EAAE,IAAI,mBAAmB;EACzD,MAAM,EAAE,YAAY;AAEpB,MAAI,aAAa;GACf;GACA,QAAQ;GACR;GACA;GACA;GACA;GACD;;AAEH,QAAO;;;;;AAMT,MAAa,uBAA2C;CACtD;CACA;CACA;CACD;;;;AAKD,SAAgB,iBACd,KACA,OACe;AACf,MAAK,MAAM,QAAQ,MAEjB,OAAM,KAAK,IAAI;AAGjB,QAAO;;;;;;;;;;;AClDT,SAAS,qBAAqB,OAAuB;AACnD,KAAI,MAAM,SAAS,SAAS,EAC1B,QAAO;CAET,MAAM,MAAM,MAAM,SAAS;AAC3B,KAAI,OAAO,IAAI,SAAS,SAAS,EAC/B,QAAO;CAET,MAAM,OAAO,KAAK,SAAS;AAC3B,QAAO,SAAS,KAAA,KAAa,KAAK,SAAS,SAAS;;AAGtD,SAAS,mBAAmB,OAAwC;AAClE,QAAO;EACL,UAAU;GAAE,KAAK;GAAO;GAAO;EAC/B,MAAM;EACN,OAAO;EACR;;AAGH,SAAS,gBAAgB,OAAqC;AAC5D,QAAO;EACL,UAAU,EAAE;EACZ,UAAU;GAAE,KAAK;GAAO;GAAO;EAC/B,MAAM;EACP;;AAGH,SAAS,sBAAsB,OAA2C;AACxE,QAAO;EACL,UAAU,EAAE;EACZ,UAAU;GAAE,KAAK;GAAO;GAAO;EAC/B,MAAM;EACP;;AAGH,SAAS,YAAY,OAAiC;AACpD,QAAO;EACL,UAAU,EAAE;EACZ,UAAU;GAAE,KAAK;GAAO;GAAO;EAC/B,MAAM;EACP;;AAIH,SAAS,mBAAmB;CAC1B,MAAM,OAAa;EACjB,UAAU,EAAE;EACZ,MAAM;EACP;CAED,IAAI,MAAsB;CAC1B,IAAI,QAAsB;CAC1B,IAAI,MAA8B;CAClC,IAAI,OAAyB;CAC7B,IAAI,aAAkC;CACtC,IAAI,oBAAoB;CACxB,IAAI,iBAAyC;CAC7C,IAAI,cAAsC;CAC1C,IAAI,uBAAuB;CAC3B,IAAI,qBAAqB;CACzB,IAAI,eAAe;CAEnB,MAAM,mBAAmB;AACvB,QAAM;AACN,UAAQ;AACR,QAAM;AACN,SAAO;AACP,eAAa;AACb,sBAAoB;AACpB,mBAAiB;AACjB,yBAAuB;;CAGzB,MAAM,eAAe,MAAc,aAAuB;AACxD,QAAM;GACJ,UAAU,EAAE;GACZ;GACA,UAAU;IAAE,KAAK,SAAS;IAAK,OAAO,SAAS;IAAO;GACtD,MAAM;GACP;AACD,OAAK,SAAS,KAAK,IAAI;AACvB,UAAQ;AACR,QAAM;AACN,SAAO;AACP,eAAa;AACb,sBAAoB;AACpB,uBAAqB;AACrB,yBAAuB;;;;;;CAOzB,MAAM,iBAAiB,UAA6B;AAClD,MAAI,CAAC,KAAK;AACR,eAAY,IAAI;IAAE,KAAK;IAAO;IAAO,CAAC;AAGtC,wBAAqB;;;CAIzB,MAAM,aAAa,UAA6B;AAC9C,gBAAc,MAAM;AACpB,UAAQ,YAAY,MAAM;AAC1B,MAAK,SAAS,KAAK,MAAM;AACzB,QAAM;AACN,SAAO;AACP,eAAa;AACb,sBAAoB;;CAGtB,MAAM,kBAAkB,UAA6B;AACnD,MAAI,CAAC,OAAO;AACV,aAAU,MAAM;GAEhB,MAAM,WAAW,sBAAsB,MAAM;AAC7C,SAAO,SAAS,KAAK,SAAS;;AAGhC,MAAI,MAAO,SAAS,WAAW,GAAG;GAChC,MAAM,WAAW,sBAAsB,MAAM;AAC7C,SAAO,SAAS,KAAK,SAAS;;AAEhC,QAAM,sBAAsB,MAAM;AAClC,QAAO,SAAS,KAAK,IAAI;AACzB,SAAO;AACP,eAAa;AACb,sBAAoB;;CAKtB,MAAM,mBAAmB,UAA6B;AACpD,MAAI,CAAC,MACH,WAAU,MAAM;AAElB,MAAI,CAAC,KAAK;AACR,SAAM,sBAAsB,MAAM;AAClC,SAAO,SAAS,KAAK,IAAI;;AAE3B,MAAI,CAAC,MAAM;AACT,UAAO,gBAAgB,MAAM;AAC7B,OAAI,SAAS,KAAK,KAAK;;AAEzB,sBAAoB;;CAGtB,MAAM,iBAAiB,UAA6B;AAClD,MAAI,CAAC,OAAO;AACV,aAAU,MAAM;AAEhB,SAAM,sBAAsB,MAAM;AAClC,SAAO,SAAS,KAAK,IAAI;GACzB,MAAM,YAAY,gBAAgB,MAAM;AACxC,OAAI,SAAS,KAAK,UAAU;;AAE9B,MAAI,CAAC,KAAK;AACR,SAAM,sBAAsB,MAAM;AAClC,SAAO,SAAS,KAAK,IAAI;GACzB,MAAM,YAAY,gBAAgB,MAAM;AACxC,OAAI,SAAS,KAAK,UAAU;aACnB,IAAI,SAAS,WAAW,GAAG;GACpC,MAAM,YAAY,gBAAgB,MAAM;AACxC,OAAI,SAAS,KAAK,UAAU;;AAE9B,SAAO,gBAAgB,MAAM;AAC7B,MAAI,SAAS,KAAK,KAAK;AACvB,eAAa;AACb,sBAAoB;;CAGtB,MAAM,iBAAiB,UAA6B;AAClD,MAAI,CAAC,MACH,WAAU,MAAM;AAElB,MAAI,CAAC,KAAK;AACR,SAAM,sBAAsB,MAAM;AAClC,SAAO,SAAS,KAAK,IAAI;;AAE3B,MAAI,CAAC,MAAM;AACT,UAAO,gBAAgB,MAAM;AAC7B,OAAI,SAAS,KAAK,KAAK;;AAEzB,MAAI,CAAC,YAAY;AACf,gBAAa,mBAAmB,MAAM;AACtC,QAAK,SAAS,KAAK,WAAW;AAC9B,uBAAoB;;;CAIxB,MAAM,wBAAwB,WAA4B;AACxD,MAAI,YAAY,SACd,YAAW,SAAS,MAAM;AAE5B,MAAI,MAAM,SACR,MAAK,SAAS,MAAM;AAEtB,MAAI,KAAK,SACP,KAAI,SAAS,MAAM;AAErB,MAAI,OAAO,SACT,OAAM,SAAS,MAAM;AAEvB,MAAI,KAAK,SACP,KAAI,SAAS,MAAM;;CAIvB,MAAM,gBAAgB,QAAe;AACnC,UAAQ,IAAI,MAAZ;GACE,KAAK;AAEH,yBADe,kBAAkB,IAAI,SAAS,MAClB;AAC5B,qCAAiC;AACjC,gBAAY;AACZ;GAEF,KAAK;AACH,qBAAiB,IAAI,SAAS;AAC9B,kBAAc,IAAI,SAAS;AAE3B,QAAI,oBAAoB;AACtB,0BAAqB;AACrB;;AAGF,QAAI,CAAC,MACH,WAAU,IAAI,SAAS,MAAM;AAE/B,cAAU,IAAI,SAAS,IAAI;AAC3B;GAEF,KAAK;AACH,qBAAiB,IAAI,SAAS;AAC9B,kBAAc,IAAI,SAAS;AAC3B,QAAI,CAAC,MACH,WAAU,IAAI,SAAS,MAAM;AAE/B,mBAAe,IAAI,SAAS,IAAI;AAChC;GAEF,KAAK;AACH,qBAAiB,IAAI,SAAS;AAC9B,kBAAc,IAAI,SAAS;AAC3B,kBAAc,IAAI,SAAS,IAAI;AAC/B;GAEF,KAAK;AACH,qBAAiB,IAAI,SAAS;AAC9B,kBAAc,IAAI,SAAS;AAC3B,oBAAgB,IAAI,SAAS,MAAM;AAEnC,QAAI,KAAM,SAAS,WAAW,GAAG;KAC/B,MAAM,WAAW,mBAAmB,IAAI,SAAS,MAAM;AACvD,UAAM,SAAS,KAAK,SAAS;;AAE/B,iBAAa,mBAAmB,IAAI,SAAS,IAAI;AACjD,UAAM,SAAS,KAAK,WAAW;AAC/B,wBAAoB;AACpB;GAEF,KAAK,QAAQ;IACX,MAAM,MAAM,IAAI,SAAS;AAGzB,QAAI,CAAC,cAAc;AACjB,UAAK,WAAW;MACd,KAAK,IAAI,SAAS;MAClB,OAAO,IAAI,SAAS;MACrB;AACD,oBAAe;;AAIjB,QAAI,sBAAsB;AACxB,iBAAY,KAAK,IAAI,SAAS;AAC9B,sBAAiB,IAAI,SAAS;AAC9B,mBAAc,IAAI,SAAS;AAC3B;;AAIF,kBAAc,IAAI,SAAS,MAAM;AACjC,QAAI,WACF,YAAW,SAAS;AAEtB,yBAAqB,IAAI,SAAS,IAAI;AACtC,qBAAiB,IAAI,SAAS;AAC9B,kBAAc,IAAI,SAAS;AAC3B,wBAAoB;AACpB;;GAEF,QACE,OAAM,IAAI,MAAM,0BAA0B,IAAI,OAAO;;;CAO3D,MAAM,iBAAiB;AACrB,MAAI,kBAAkB,IACpB,sBAAqB,eAAe;AAEtC,mCAAiC;AACjC,MAAI,OAAO,CAAC,qBAAqB,CAAC,IAAI,KASpC,MAAK,SAAS,KAAK;AAIrB,MAAI,KAAK,YAAY,YACnB,MAAK,SAAS,MAAM;WACX,CAAC,KAAK,SACf,MAAK,WAAW;GACd,KAAK;IAAE,QAAQ;IAAG,MAAM;IAAG,QAAQ;IAAG;GACtC,OAAO;IAAE,QAAQ;IAAG,MAAM;IAAG,QAAQ;IAAG;GACzC;AAGH,SAAO;;;;;;;CAQT,SAAS,kCAAkC;AACzC,MAAI,CAAC,OAAO,IAAI,SAAS,WAAW,EAClC;EAEF,MAAM,YAAY,IAAI,SAAS,GAAG,GAAG;AACrC,MAAI,WAAW,SAAS,QACtB;EAEF,MAAM,YAAY;AAGlB,MAAI,YAAY,UAAU,IAAI,CAAC,qBAAqB,UAAU,CAC5D,KAAI,SAAS,KAAK;;AAItB,QAAO;EAAE;EAAU;EAAc;EAAM;;AAGzC,SAAgB,uBAAuB,QAA+B;CACpE,MAAM,OAAO,kBAAkB;AAC/B,MAAK,MAAM,OAAO,OAChB,MAAK,aAAa,IAAI;AAExB,QAAO,KAAK,UAAU;;;;ACxXxB,MAAM,oBAAoB;AAC1B,MAAM,kBAAkB;AACxB,MAAM,4BAA4B;AAClC,MAAM,0BAA0B;AAChC,MAAM,4BAA4B;AAClC,MAAM,0BAA0B;AAEhC,IAAa,iBAAb,MAAkE;CAChE,QAAgB;CAChB,IAAY;CACZ,OAAe;CACf,MAAc;CACd;CAGA,kBAA0B;CAC1B,mBAGI;CAEJ,MAAM,KAAoB;AACxB,OAAK,QAAQ,IAAI;AACjB,OAAK,SAAS,IAAI;AAClB,OAAK,IAAI;AACT,OAAK,OAAO;AACZ,OAAK,MAAM;AACX,OAAK,kBAAkB;AACvB,OAAK,mBAAmB;AAGxB,MAAI,KAAK,MAAM,WAAW,MAAM,EAAE;GAEhC,MAAM,MAAM,KAAK,OAAO,mBAAmB,gBAAgB;GAC3D,MAAM,OAAO,KAAK,OAChB,2BACA,wBACD;GACD,MAAM,OAAO,KAAK,OAChB,2BACA,wBACD;AACD,QAAK,mBAAmB;IACtB;KAAE,SAAS,IAAI;KAAQ,MAAM;KAAQ,OAAO;KAAK;IACjD;KAAE,SAAS;KAAG,MAAM;KAAe;IACnC;KAAE,SAAS,KAAK;KAAQ,MAAM;KAAQ,OAAO;KAAM;IACnD;KAAE,SAAS;KAAG,MAAM;KAAe;IACnC;KAAE,SAAS,KAAK;KAAQ,MAAM;KAAQ,OAAO;KAAM;IACpD;;;CAML,OAAqB;EACnB,MAAM,IAAI,KAAK;EACf,MAAM,IAAI,EAAE;AACZ,MAAI,KAAK,KAAK,KAAK,CAAC,KAAK,kBAAkB,OACzC,QAAO;EAGT,MAAM,QAAQ;GAAE,QAAQ,KAAK;GAAK,MAAM,KAAK;GAAM,QAAQ,KAAK;GAAG;AAGnE,MAAI,CAAC,KAAK,mBAAmB,KAAK,kBAAkB;GAClD,MAAM,OAAO,KAAK,iBAAiB,OAAO;AAC1C,OAAI,MAAM;AACR,QAAI,KAAK,SAAS,QAAQ;KACxB,MAAM,OAAO,KAAK,IAAI,KAAK,SAAS,IAAI,KAAK,EAAE;KAC/C,MAAM,MAAM,KAAK,MAAM,MAAM,GAAG,KAAK;AACrC,UAAK,aAAa,IAAI;AACtB,SAAI,KAAK,iBAAiB,WAAW,GAAG;AACtC,WAAK,mBAAmB;AACxB,WAAK,kBAAkB;;AAEzB,YAAO,KAAK,KAAK,QAAQ,KAAK,MAAM;;AAGtC,SAAK,SAAS,KAAK,IAAI,KAAK,SAAS,IAAI,KAAK,EAAE,CAAC;AACjD,QAAI,KAAK,iBAAiB,WAAW,GAAG;AACtC,UAAK,mBAAmB;AACxB,UAAK,kBAAkB;;AAEzB,WAAO,KAAK,KAAK,eAAe,KAAA,GAAW,MAAM;;;AAKrD,MAAI,KAAK,KAAK,EACZ,QAAO;EAET,MAAM,KAAK,EAAE,OAAO,KAAK,EAAE;AAE3B,MAAI,OAAO,KAAK,OAAO,SAAS;AAC9B,QAAK,SAAS,GAAG,KAAK;AACtB,UAAO,KAAK,KAAK,eAAe,KAAA,GAAW,MAAM;;AAEnD,MAAI,OAAO,KAAK,OAAO,OAAO;AAC5B,QAAK,SAAS,EAAE;AAChB,UAAO,KAAK,KAAK,eAAe,KAAA,GAAW,MAAM;;AAEnD,MAAI,OAAO,KAAK,OAAO,YAAY;AACjC,QAAK,SAAS,EAAE;AAChB,UAAO,KAAK,KAAK,oBAAoB,KAAA,GAAW,MAAM;;AAExD,MAAI,OAAO,KAAK,OAAO,WAAW;AAChC,QAAK,SAAS,EAAE;AAChB,UAAO,KAAK,KAAK,mBAAmB,KAAA,GAAW,MAAM;;AAEvD,MAAI,OAAO,KAAK,OAAO,cAAc;AACnC,QAAK,SAAS,EAAE;AAChB,UAAO,KAAK,KAAK,iBAAiB,KAAA,GAAW,MAAM;;EAIrD,IAAI,IAAI,KAAK;EACb,MAAM,MAAM,KAAK,OAAO,WACtB,MAAM,KAAK,OAAO,OAClB,MAAM,KAAK,OAAO,YAClB,MAAM,KAAK,OAAO,SAClB,MAAM,KAAK,OAAO;AAEpB,SAAO,IAAI,EAAE,QAAQ;GACnB,MAAM,IAAI,EAAE,OAAO,EAAE;AACrB,OAAI,MAAM,OAAO,MAAM,OAAO,MAAM,OAAO,MAAM,OAAO,MAAM,IAC5D;AAEF,QAAK;;EAGP,MAAM,MAAM,EAAE,MAAM,KAAK,GAAG,EAAE;AAC9B,OAAK,aAAa,IAAI;AACtB,SAAO,KAAK,KAAK,QAAQ,KAAK,MAAM;;CAItC,OAAe,OAAe,KAAqB;AACjD,SAAO,KAAK,MAAM,MAAM,OAAO,KAAK,IAAI,KAAK,KAAK,MAAM,OAAO,CAAC;;CAGlE,SAAiB,GAAW,YAAY,OAAO;AAC7C,OAAK,KAAK;AACV,MAAI,WAAW;AACb,QAAK,QAAQ;AACb,QAAK,MAAM;QAEX,MAAK,OAAO;;CAIhB,aAAqB,OAAe;EAClC,MAAM,QAAQ,MAAM,MAAM,KAAK,OAAO,QAAQ;AAC9C,MAAI,MAAM,SAAS,GAAG;AACpB,QAAK,QAAQ,MAAM,SAAS;AAC5B,QAAK,OAAO,MAAM,GAAG,GAAG,EAAE,UAAU,KAAK;QAEzC,MAAK,OAAO,MAAM;AAEpB,OAAK,KAAK,MAAM;;CAGlB,KACE,MACA,OACA,OACO;AAEP,SAAO;GAAE,UAAU;IAAE,KAAA;KADP,QAAQ,KAAK;KAAK,MAAM,KAAK;KAAM,QAAQ,KAAK;KACtC;IAAE;IAAO;GAAE;GAAM;GAAO;;CAIlD,CAAC,OAAO,YAA6B;AACnC,SAAO,EACL,YAAY;GACV,MAAM,IAAI,KAAK,MAAM;AACrB,OAAI,CAAC,EACH,QAAO;IAAE,MAAM;IAAM,OAAO,KAAA;IAA+B;AAE7D,UAAO;IAAE,MAAM;IAAO,OAAO;IAAG;KAEnC;;;;;AChLL,UAAU,qBAAqB,GAA+B;AAC5D,MAAK,IAAI,MAAM,EAAE,MAAM,EAAE,KAAK,MAAM,EAAE,MAAM,CAC1C,OAAM;;AAIV,SAAgB,WACd,OACA,OAAqB,EAAE,EACvB,UACM;CAKN,IAAI,MAAqB;EACvB,YAAY;GACV,GAAG;GACH,GAAI,UAAU,cAAc,EAAE;GAC/B;EACD;EACD;AAED,OAAM,iBAAiB,KAAK,KAAK,cAAc,qBAAqB;CAEpE,MAAM,YAAY,IAAI,gBAAgB;AAEtC,WAAU,MAAM,IAAI;AAEpB,QAAO,uBAAuB,qBAAqB,UAAU,CAAC;;AAGhE,MAAM,cAAqD,SAAS,YAElE,UAAU,EAAE,EACZ;AACA,MAAK,UAAU,aAAqB;AAIlC,SAAO,WAAW,UAAU,SAFX,KAAK,KAAK,WAEkB,CAAC"}
+{"version":3,"file":"index.js","names":[],"sources":["../src/preprocessor.ts","../src/processor.ts","../src/tokenizer.ts","../src/parser.ts"],"sourcesContent":["import { DEFAULT_DELIMITERS } from \"@glion/utils\";\n\nimport type { ParserContext, PreprocessorStep } from \"./types\";\n\n// PreprocessorStep is declared in `types.ts` to avoid circular imports\n\n/**\n * Step: strip UTF-8 BOM.\n */\nexport const stripBOM: PreprocessorStep = (ctx) => {\n  if (ctx.input.codePointAt(0) === 0xfe_ff) {\n    ctx.input = ctx.input.slice(1);\n  }\n  return ctx;\n};\n\n/**\n * Step: normalize newlines to the default delimiters.\n */\nexport const normalizeNewlines: PreprocessorStep = (ctx) => {\n  ctx.input = ctx.input.replaceAll(/\\r?\\n/g, ctx.delimiters.segment);\n  return ctx;\n};\n\n/**\n * Step: auto-detect delimiters from MSH.\n */\nexport const detectDelimiters: PreprocessorStep = (ctx) => {\n  if (ctx.input.startsWith(\"MSH\")) {\n    const field = ctx.input.charAt(3) || DEFAULT_DELIMITERS.field;\n    const enc = ctx.input.slice(4, 8);\n    const component = enc.charAt(0) || DEFAULT_DELIMITERS.component;\n    const repetition = enc.charAt(1) || DEFAULT_DELIMITERS.repetition;\n    const _escape = enc.charAt(2) || DEFAULT_DELIMITERS.escape;\n    const subcomponent = enc.charAt(3) || DEFAULT_DELIMITERS.subcomponent;\n    const { segment } = DEFAULT_DELIMITERS;\n\n    ctx.delimiters = {\n      component,\n      escape: _escape,\n      field,\n      repetition,\n      segment,\n      subcomponent,\n    };\n  }\n  return ctx;\n};\n\n/**\n * Default preprocessing pipeline.\n */\nexport const defaultPreprocessors: PreprocessorStep[] = [\n  stripBOM,\n  normalizeNewlines,\n  detectDelimiters,\n];\n\n/**\n * Run the pipeline to initialize ParserContext.\n */\nexport function runPreprocessors(\n  ctx: ParserContext,\n  steps: PreprocessorStep[]\n): ParserContext {\n  for (const step of steps) {\n    // oxlint-disable-next-line no-param-reassign\n    ctx = step(ctx);\n  }\n\n  return ctx;\n}\n","import type {\n  Component,\n  Field,\n  FieldRepetition,\n  Root,\n  Segment,\n  Subcomponent,\n} from \"@glion/ast\";\nimport { isEmptyNode } from \"@glion/utils\";\nimport type { Position } from \"unist\";\n\nimport type { Token } from \"./types\";\n\n/**\n * Check if a field has structural content beyond a single empty value.\n * A field with multiple repetitions, multiple components, or multiple\n * subcomponents has structural content even if all values are empty.\n * This is used to prevent dropping trailing fields like |^| or |&|\n * that have meaningful delimiter structure.\n */\nfunction hasStructuralContent(field: Field): boolean {\n  if (field.children.length > 1) {\n    return true;\n  }\n  const rep = field.children[0];\n  if (rep && rep.children.length > 1) {\n    return true;\n  }\n  const comp = rep?.children[0];\n  return comp !== undefined && comp.children.length > 1;\n}\n\nfunction createSubcomponent(start: Position[\"start\"]): Subcomponent {\n  return {\n    position: { end: start, start },\n    type: \"subcomponent\",\n    value: \"\",\n  };\n}\n\nfunction createComponent(start: Position[\"start\"]): Component {\n  return {\n    children: [],\n    position: { end: start, start },\n    type: \"component\",\n  };\n}\n\nfunction createFieldRepetition(start: Position[\"start\"]): FieldRepetition {\n  return {\n    children: [],\n    position: { end: start, start },\n    type: \"field-repetition\",\n  };\n}\n\nfunction createField(start: Position[\"start\"]): Field {\n  return {\n    children: [],\n    position: { end: start, start },\n    type: \"field\",\n  };\n}\n\n// Shared core: process a single token into mutable parse state\nfunction createParserCore() {\n  const root: Root = {\n    children: [],\n    type: \"root\",\n  };\n\n  let seg: Segment | null = null;\n  let field: Field | null = null;\n  let rep: FieldRepetition | null = null;\n  let comp: Component | null = null;\n  let currentSub: Subcomponent | null = null;\n  let segmentHasContent = false;\n  let lastContentEnd: Position[\"end\"] | null = null;\n  let documentEnd: Position[\"end\"] | null = null; // Track document-level end position\n  let expectingSegmentName = true; // Start expecting a segment name\n  let justSetSegmentName = false; // Track if we just set the segment name\n  let rootStartSet = false; // Track if we've set the root start position\n\n  const resetState = () => {\n    seg = null;\n    field = null;\n    rep = null;\n    comp = null;\n    currentSub = null;\n    segmentHasContent = false;\n    lastContentEnd = null;\n    expectingSegmentName = true;\n  };\n\n  const openSegment = (name: string, position: Position) => {\n    seg = {\n      children: [],\n      name,\n      position: { end: position.end, start: position.start },\n      type: \"segment\",\n    };\n    root.children.push(seg);\n    field = null;\n    rep = null;\n    comp = null;\n    currentSub = null;\n    segmentHasContent = false;\n    justSetSegmentName = true;\n    expectingSegmentName = false;\n  };\n\n  /**\n   * Ensure there is an active segment. If delimiter tokens arrive before\n   * any segment name, lazily open a segment with an empty name.\n   */\n  const ensureSegment = (start: Position[\"start\"]) => {\n    if (!seg) {\n      openSegment(\"\", { end: start, start });\n      // Reset justSetSegmentName so the first FIELD_DELIM isn't skipped\n      // (there's no real segment name to separate from fields)\n      justSetSegmentName = false;\n    }\n  };\n\n  const openField = (start: Position[\"start\"]) => {\n    ensureSegment(start);\n    field = createField(start);\n    seg!.children.push(field);\n    rep = null;\n    comp = null;\n    currentSub = null;\n    segmentHasContent = true;\n  };\n\n  const openRepetition = (start: Position[\"start\"]) => {\n    if (!field) {\n      openField(start);\n      // Add an empty first repetition before the delimiter\n      const emptyRep = createFieldRepetition(start);\n      field!.children.push(emptyRep);\n    }\n    // If field has no children yet, add empty first rep\n    if (field!.children.length === 0) {\n      const emptyRep = createFieldRepetition(start);\n      field!.children.push(emptyRep);\n    }\n    rep = createFieldRepetition(start);\n    field!.children.push(rep);\n    comp = null;\n    currentSub = null;\n    segmentHasContent = true;\n  };\n\n  // Ensure there's a component to add subcomponents to (for SUBCOMP_DELIM)\n  // Does NOT add empty sibling components.\n  const ensureComponent = (start: Position[\"start\"]) => {\n    if (!field) {\n      openField(start);\n    }\n    if (!rep) {\n      rep = createFieldRepetition(start);\n      field!.children.push(rep);\n    }\n    if (!comp) {\n      comp = createComponent(start);\n      rep.children.push(comp);\n    }\n    segmentHasContent = true;\n  };\n\n  const openComponent = (start: Position[\"start\"]) => {\n    if (!field) {\n      openField(start);\n      // Build out a rep + empty first component before the delimiter\n      rep = createFieldRepetition(start);\n      field!.children.push(rep);\n      const emptyComp = createComponent(start);\n      rep.children.push(emptyComp);\n    }\n    if (!rep) {\n      rep = createFieldRepetition(start);\n      field!.children.push(rep);\n      const emptyComp = createComponent(start);\n      rep.children.push(emptyComp);\n    } else if (rep.children.length === 0) {\n      const emptyComp = createComponent(start);\n      rep.children.push(emptyComp);\n    }\n    comp = createComponent(start);\n    rep.children.push(comp);\n    currentSub = null;\n    segmentHasContent = true;\n  };\n\n  const ensureForText = (start: Position[\"start\"]) => {\n    if (!field) {\n      openField(start);\n    }\n    if (!rep) {\n      rep = createFieldRepetition(start);\n      field!.children.push(rep);\n    }\n    if (!comp) {\n      comp = createComponent(start);\n      rep.children.push(comp);\n    }\n    if (!currentSub) {\n      currentSub = createSubcomponent(start);\n      comp.children.push(currentSub);\n      segmentHasContent = true;\n    }\n  };\n\n  const updatePositionsToEnd = (endPos: Position[\"end\"]) => {\n    if (currentSub?.position) {\n      currentSub.position.end = endPos;\n    }\n    if (comp?.position) {\n      comp.position.end = endPos;\n    }\n    if (rep?.position) {\n      rep.position.end = endPos;\n    }\n    if (field?.position) {\n      field.position.end = endPos;\n    }\n    if (seg?.position) {\n      seg.position.end = endPos;\n    }\n  };\n\n  const processToken = (tok: Token) => {\n    switch (tok.type) {\n      case \"SEGMENT_END\": {\n        const endPos = lastContentEnd || tok.position.start;\n        updatePositionsToEnd(endPos);\n        dropTrailingEmptyFieldIfPresent();\n        resetState();\n        return;\n      }\n      case \"FIELD_DELIM\": {\n        lastContentEnd = tok.position.end;\n        documentEnd = tok.position.end;\n        // Skip the first field delimiter after setting segment name\n        if (justSetSegmentName) {\n          justSetSegmentName = false;\n          return;\n        }\n        // Leading field delimiter implies an empty first field\n        if (!field) {\n          openField(tok.position.start);\n        }\n        openField(tok.position.end);\n        return;\n      }\n      case \"REPETITION_DELIM\": {\n        lastContentEnd = tok.position.end;\n        documentEnd = tok.position.end;\n        if (!field) {\n          openField(tok.position.start);\n        }\n        openRepetition(tok.position.end);\n        return;\n      }\n      case \"COMPONENT_DELIM\": {\n        lastContentEnd = tok.position.end;\n        documentEnd = tok.position.end;\n        openComponent(tok.position.end);\n        return;\n      }\n      case \"SUBCOMP_DELIM\": {\n        lastContentEnd = tok.position.end;\n        documentEnd = tok.position.end;\n        ensureComponent(tok.position.start);\n        // If comp has no children yet, add empty first subcomponent\n        if (comp!.children.length === 0) {\n          const emptySub = createSubcomponent(tok.position.start);\n          comp!.children.push(emptySub);\n        }\n        currentSub = createSubcomponent(tok.position.end);\n        comp?.children.push(currentSub);\n        segmentHasContent = true;\n        return;\n      }\n      case \"TEXT\": {\n        const val = tok.value ?? \"\";\n\n        // Set root start position on first token\n        if (!rootStartSet) {\n          root.position = {\n            end: tok.position.start,\n            start: tok.position.start,\n          };\n          rootStartSet = true;\n        }\n\n        // If we're expecting a segment name, this TEXT is the segment name\n        if (expectingSegmentName) {\n          openSegment(val, tok.position);\n          lastContentEnd = tok.position.end;\n          documentEnd = tok.position.end;\n          return;\n        }\n\n        // Otherwise, it's regular field content\n        ensureForText(tok.position.start);\n        if (currentSub) {\n          currentSub.value += val;\n        }\n        updatePositionsToEnd(tok.position.end);\n        lastContentEnd = tok.position.end;\n        documentEnd = tok.position.end;\n        segmentHasContent = true;\n        return;\n      }\n      default: {\n        throw new Error(`Unexpected token type: ${tok.type}`);\n      }\n    }\n  };\n\n  // Do not pre-open a segment; lazily open upon first non-SEGMENT_END token.\n\n  const finalize = () => {\n    if (lastContentEnd && seg) {\n      updatePositionsToEnd(lastContentEnd);\n    }\n    dropTrailingEmptyFieldIfPresent();\n    if (seg && !segmentHasContent && !seg.name) {\n      // Drop trailing empty segment only when it was lazily opened by\n      // `ensureSegment` (no name, no content). A segment that arrived\n      // with a name carries information — even if no fields followed —\n      // and dropping it would break round-trip idempotency: the\n      // serializer re-emits the name, the next parse would drop it\n      // again, and any sibling segment immediately preceding it would\n      // shift into the trailing position and get dropped on the\n      // following parse.\n      root.children.pop();\n    }\n\n    // Set root end position once at finalization using document-level end\n    if (root.position && documentEnd) {\n      root.position.end = documentEnd;\n    } else if (!root.position) {\n      root.position = {\n        end: { column: 1, line: 1, offset: 0 },\n        start: { column: 1, line: 1, offset: 0 },\n      };\n    }\n\n    return root;\n  };\n\n  /**\n   * Drop only the phantom trailing empty field created by a terminal field\n   * delimiter (e.g. the last `|` in `PID|foo|`). Without this, every segment\n   * ending with `|` would gain an extra empty field on each parse cycle.\n   */\n  function dropTrailingEmptyFieldIfPresent() {\n    if (!seg || seg.children.length === 0) {\n      return;\n    }\n    const lastChild = seg.children.at(-1);\n    if (lastChild?.type !== \"field\") {\n      return;\n    }\n    const lastField = lastChild as Field;\n    // A field like `|^|` is isEmpty (no leaf data) but has structural content\n    // (component separators) and must be preserved.\n    if (isEmptyNode(lastField) && !hasStructuralContent(lastField)) {\n      seg.children.pop();\n    }\n  }\n\n  return { finalize, processToken, root };\n}\n\nexport function parseHL7v2FromIterator(tokens: Iterable<Token>): Root {\n  const core = createParserCore();\n  for (const tok of tokens) {\n    core.processToken(tok);\n  }\n  return core.finalize();\n}\n","// src/tokenizer.ts\nimport type { Delimiters } from \"@glion/ast\";\nimport type { Position } from \"unist\";\n\nimport type { ParserContext, Token, Tokenizer, TokenType } from \"./types\";\n\nconst MSH_SEGMENT_START = 0;\nconst MSH_SEGMENT_END = 3;\nconst MSH_FIELD_SEPERATOR_START = 3;\nconst MSH_FIELD_SEPERATOR_END = 4;\nconst MSH_FIELD_DELIMITER_START = 4;\nconst MSH_FIELD_DELIMITER_END = 8;\n\nexport class HL7v2Tokenizer implements Tokenizer, Iterable<Token> {\n  private input = \"\";\n  private i = 0;\n  private line = 1;\n  private col = 1;\n  private delims!: Delimiters;\n\n  // Only-run-once MSH bootstrap at the start of the file\n  private didMshBootstrap = false;\n  private pendingBootstrap: null | Array<\n    | { kind: \"TEXT\"; value: string; advance: number }\n    | { kind: \"FIELD_DELIM\"; advance: number }\n  > = null; // queue to emit TEXT('MSH'), FIELD_DELIM, TEXT('^~\\\\&')\n\n  reset(ctx: ParserContext) {\n    this.input = ctx.input;\n    this.delims = ctx.delimiters;\n    this.i = 0;\n    this.line = 1;\n    this.col = 1;\n    this.didMshBootstrap = false;\n    this.pendingBootstrap = null;\n\n    // Prepare a one-time bootstrap if file starts with MSH\n    if (this.input.startsWith(\"MSH\")) {\n      // Precompute MSH and MSH.2; MSH.1 is the field delimiter char at index 3\n      const msh = this._slice(MSH_SEGMENT_START, MSH_SEGMENT_END);\n      const msh1 = this._slice(\n        MSH_FIELD_SEPERATOR_START,\n        MSH_FIELD_SEPERATOR_END\n      ); // the field delimiter char at index 3\n      const msh2 = this._slice(\n        MSH_FIELD_DELIMITER_START,\n        MSH_FIELD_DELIMITER_END\n      ); // may be shorter than 4 if truncated\n      this.pendingBootstrap = [\n        { advance: msh.length, kind: \"TEXT\", value: msh },\n        { advance: 0, kind: \"FIELD_DELIM\" },\n        { advance: msh1.length, kind: \"TEXT\", value: msh1 },\n        { advance: 0, kind: \"FIELD_DELIM\" },\n        { advance: msh2.length, kind: \"TEXT\", value: msh2 },\n      ];\n      // NOTE: we have not advanced indices yet; we will as we emit tokens\n    }\n  }\n\n  // oxlint-disable-next-line complexity/complexity -- tokenizer hot path: one branch per delimiter class, kept flat for speed\n  next(): Token | null {\n    const s = this.input;\n    const n = s.length;\n    if (this.i >= n && !this.pendingBootstrap?.length) {\n      return null;\n    }\n\n    const start = { column: this.col, line: this.line, offset: this.i };\n\n    // ---- One-time MSH bootstrap at file start ----\n    if (!this.didMshBootstrap && this.pendingBootstrap) {\n      const step = this.pendingBootstrap.shift();\n      if (step) {\n        if (step.kind === \"TEXT\") {\n          const take = Math.min(step.advance, n - this.i);\n          const out = step.value.slice(0, take);\n          this._fastAdvance(out);\n          if (this.pendingBootstrap.length === 0) {\n            this.pendingBootstrap = null;\n            this.didMshBootstrap = true;\n          }\n          return this._tok(\"TEXT\", out, start);\n        }\n        // FIELD_DELIM: advance exactly one char (the delimiter) and emit a FIELD_DELIM token\n        this._advance(Math.min(step.advance, n - this.i));\n        if (this.pendingBootstrap.length === 0) {\n          this.pendingBootstrap = null;\n          this.didMshBootstrap = true;\n        }\n        return this._tok(\"FIELD_DELIM\", undefined, start);\n      }\n    }\n\n    // ---- Normal tokenization (delimiters + text) ----\n    if (this.i >= n) {\n      return null;\n    }\n    const ch = s.charAt(this.i);\n\n    if (ch === this.delims.segment) {\n      this._advance(1, true);\n      return this._tok(\"SEGMENT_END\", undefined, start);\n    }\n    if (ch === this.delims.field) {\n      this._advance(1);\n      return this._tok(\"FIELD_DELIM\", undefined, start);\n    }\n    if (ch === this.delims.repetition) {\n      this._advance(1);\n      return this._tok(\"REPETITION_DELIM\", undefined, start);\n    }\n    if (ch === this.delims.component) {\n      this._advance(1);\n      return this._tok(\"COMPONENT_DELIM\", undefined, start);\n    }\n    if (ch === this.delims.subcomponent) {\n      this._advance(1);\n      return this._tok(\"SUBCOMP_DELIM\", undefined, start);\n    }\n\n    // TEXT until next delimiter or end\n    let j = this.i;\n    const cmp = this.delims.component,\n      fld = this.delims.field,\n      rep = this.delims.repetition,\n      seg = this.delims.segment,\n      sub = this.delims.subcomponent;\n\n    while (j < s.length) {\n      const c = s.charAt(j);\n      if (c === seg || c === fld || c === rep || c === cmp || c === sub) {\n        break;\n      }\n      j += 1;\n    }\n\n    const val = s.slice(this.i, j);\n    this._fastAdvance(val);\n    return this._tok(\"TEXT\", val, start);\n  }\n\n  // ---- helpers ----\n  private _slice(start: number, end: number): string {\n    return this.input.slice(start, Math.min(end, this.input.length));\n  }\n\n  private _advance(n: number, isNewline = false) {\n    this.i += n;\n    if (isNewline) {\n      this.line += 1;\n      this.col = 1;\n    } else {\n      this.col += n;\n    }\n  }\n\n  private _fastAdvance(chunk: string) {\n    const parts = chunk.split(this.delims.segment);\n    if (parts.length > 1) {\n      this.line += parts.length - 1;\n      this.col = (parts.at(-1)?.length ?? 0) + 1;\n    } else {\n      this.col += chunk.length;\n    }\n    this.i += chunk.length;\n  }\n\n  private _tok(\n    type: TokenType,\n    value: string | undefined,\n    start: Position[\"start\"]\n  ): Token {\n    const end = { column: this.col, line: this.line, offset: this.i };\n    return { position: { end, start }, type, value };\n  }\n\n  // Iterable protocol (sync)\n  [Symbol.iterator](): Iterator<Token> {\n    return {\n      next: () => {\n        const t = this.next();\n        if (!t) {\n          return { done: true, value: undefined as unknown as Token };\n        }\n        return { done: false, value: t };\n      },\n    };\n  }\n}\n","import type { Root } from \"@glion/ast\";\nimport type { HL7v2Settings } from \"@glion/config\";\nimport { DEFAULT_DELIMITERS } from \"@glion/utils\";\nimport type { Plugin, Processor } from \"unified\";\n\nimport { defaultPreprocessors, runPreprocessors } from \"./preprocessor\";\nimport { parseHL7v2FromIterator } from \"./processor\";\nimport { HL7v2Tokenizer } from \"./tokenizer\";\nimport type { ParseOptions, ParserContext, Token, Tokenizer } from \"./types\";\n\nfunction* iterateTokenizerSync(t: Tokenizer): Iterable<Token> {\n  for (let tok = t.next(); tok; tok = t.next()) {\n    yield tok;\n  }\n}\n\nexport function parseHL7v2(\n  input: string,\n  opts: ParseOptions = {},\n  settings?: HL7v2Settings\n): Root {\n  // Merge settings: defaults < provided settings\n  // Note: MSH auto-detection (via detectDelimiters preprocessor) runs after this\n  // and will override delimiters if the message starts with MSH\n\n  let ctx: ParserContext = {\n    delimiters: {\n      ...DEFAULT_DELIMITERS,\n      ...(settings?.delimiters ?? {}),\n    },\n    input,\n  };\n  // Run preprocessing\n  ctx = runPreprocessors(ctx, opts.preprocess || defaultPreprocessors);\n  // Run tokenizer\n  const tokenizer = new HL7v2Tokenizer();\n  // Reset tokenizer.\n  tokenizer.reset(ctx);\n  // Parse\n  return parseHL7v2FromIterator(iterateTokenizerSync(tokenizer));\n}\n\nconst hl7v2Parser: Plugin<[ParseOptions?], string, Root> = function hl7v2Parser(\n  this: Processor,\n  options = {}\n) {\n  this.parser = (document: string) => {\n    // Get settings from processor (populated by unified-args from config)\n    const settings = this.data(\"settings\");\n\n    return parseHL7v2(document, options, settings);\n  };\n};\n\nexport default hl7v2Parser;\n"],"mappings":";;;;;AASA,MAAa,YAA8B,QAAQ;AACjD,KAAI,IAAI,MAAM,YAAY,EAAE,KAAK,MAC/B,KAAI,QAAQ,IAAI,MAAM,MAAM,EAAE;AAEhC,QAAO;;;;;AAMT,MAAa,qBAAuC,QAAQ;AAC1D,KAAI,QAAQ,IAAI,MAAM,WAAW,UAAU,IAAI,WAAW,QAAQ;AAClE,QAAO;;;;;AAMT,MAAa,oBAAsC,QAAQ;AACzD,KAAI,IAAI,MAAM,WAAW,MAAM,EAAE;EAC/B,MAAM,QAAQ,IAAI,MAAM,OAAO,EAAE,IAAI,mBAAmB;EACxD,MAAM,MAAM,IAAI,MAAM,MAAM,GAAG,EAAE;EACjC,MAAM,YAAY,IAAI,OAAO,EAAE,IAAI,mBAAmB;EACtD,MAAM,aAAa,IAAI,OAAO,EAAE,IAAI,mBAAmB;EACvD,MAAM,UAAU,IAAI,OAAO,EAAE,IAAI,mBAAmB;EACpD,MAAM,eAAe,IAAI,OAAO,EAAE,IAAI,mBAAmB;EACzD,MAAM,EAAE,YAAY;AAEpB,MAAI,aAAa;GACf;GACA,QAAQ;GACR;GACA;GACA;GACA;GACD;;AAEH,QAAO;;;;;AAMT,MAAa,uBAA2C;CACtD;CACA;CACA;CACD;;;;AAKD,SAAgB,iBACd,KACA,OACe;AACf,MAAK,MAAM,QAAQ,MAEjB,OAAM,KAAK,IAAI;AAGjB,QAAO;;;;;;;;;;;AClDT,SAAS,qBAAqB,OAAuB;AACnD,KAAI,MAAM,SAAS,SAAS,EAC1B,QAAO;CAET,MAAM,MAAM,MAAM,SAAS;AAC3B,KAAI,OAAO,IAAI,SAAS,SAAS,EAC/B,QAAO;CAET,MAAM,OAAO,KAAK,SAAS;AAC3B,QAAO,SAAS,KAAA,KAAa,KAAK,SAAS,SAAS;;AAGtD,SAAS,mBAAmB,OAAwC;AAClE,QAAO;EACL,UAAU;GAAE,KAAK;GAAO;GAAO;EAC/B,MAAM;EACN,OAAO;EACR;;AAGH,SAAS,gBAAgB,OAAqC;AAC5D,QAAO;EACL,UAAU,EAAE;EACZ,UAAU;GAAE,KAAK;GAAO;GAAO;EAC/B,MAAM;EACP;;AAGH,SAAS,sBAAsB,OAA2C;AACxE,QAAO;EACL,UAAU,EAAE;EACZ,UAAU;GAAE,KAAK;GAAO;GAAO;EAC/B,MAAM;EACP;;AAGH,SAAS,YAAY,OAAiC;AACpD,QAAO;EACL,UAAU,EAAE;EACZ,UAAU;GAAE,KAAK;GAAO;GAAO;EAC/B,MAAM;EACP;;AAIH,SAAS,mBAAmB;CAC1B,MAAM,OAAa;EACjB,UAAU,EAAE;EACZ,MAAM;EACP;CAED,IAAI,MAAsB;CAC1B,IAAI,QAAsB;CAC1B,IAAI,MAA8B;CAClC,IAAI,OAAyB;CAC7B,IAAI,aAAkC;CACtC,IAAI,oBAAoB;CACxB,IAAI,iBAAyC;CAC7C,IAAI,cAAsC;CAC1C,IAAI,uBAAuB;CAC3B,IAAI,qBAAqB;CACzB,IAAI,eAAe;CAEnB,MAAM,mBAAmB;AACvB,QAAM;AACN,UAAQ;AACR,QAAM;AACN,SAAO;AACP,eAAa;AACb,sBAAoB;AACpB,mBAAiB;AACjB,yBAAuB;;CAGzB,MAAM,eAAe,MAAc,aAAuB;AACxD,QAAM;GACJ,UAAU,EAAE;GACZ;GACA,UAAU;IAAE,KAAK,SAAS;IAAK,OAAO,SAAS;IAAO;GACtD,MAAM;GACP;AACD,OAAK,SAAS,KAAK,IAAI;AACvB,UAAQ;AACR,QAAM;AACN,SAAO;AACP,eAAa;AACb,sBAAoB;AACpB,uBAAqB;AACrB,yBAAuB;;;;;;CAOzB,MAAM,iBAAiB,UAA6B;AAClD,MAAI,CAAC,KAAK;AACR,eAAY,IAAI;IAAE,KAAK;IAAO;IAAO,CAAC;AAGtC,wBAAqB;;;CAIzB,MAAM,aAAa,UAA6B;AAC9C,gBAAc,MAAM;AACpB,UAAQ,YAAY,MAAM;AAC1B,MAAK,SAAS,KAAK,MAAM;AACzB,QAAM;AACN,SAAO;AACP,eAAa;AACb,sBAAoB;;CAGtB,MAAM,kBAAkB,UAA6B;AACnD,MAAI,CAAC,OAAO;AACV,aAAU,MAAM;GAEhB,MAAM,WAAW,sBAAsB,MAAM;AAC7C,SAAO,SAAS,KAAK,SAAS;;AAGhC,MAAI,MAAO,SAAS,WAAW,GAAG;GAChC,MAAM,WAAW,sBAAsB,MAAM;AAC7C,SAAO,SAAS,KAAK,SAAS;;AAEhC,QAAM,sBAAsB,MAAM;AAClC,QAAO,SAAS,KAAK,IAAI;AACzB,SAAO;AACP,eAAa;AACb,sBAAoB;;CAKtB,MAAM,mBAAmB,UAA6B;AACpD,MAAI,CAAC,MACH,WAAU,MAAM;AAElB,MAAI,CAAC,KAAK;AACR,SAAM,sBAAsB,MAAM;AAClC,SAAO,SAAS,KAAK,IAAI;;AAE3B,MAAI,CAAC,MAAM;AACT,UAAO,gBAAgB,MAAM;AAC7B,OAAI,SAAS,KAAK,KAAK;;AAEzB,sBAAoB;;CAGtB,MAAM,iBAAiB,UAA6B;AAClD,MAAI,CAAC,OAAO;AACV,aAAU,MAAM;AAEhB,SAAM,sBAAsB,MAAM;AAClC,SAAO,SAAS,KAAK,IAAI;GACzB,MAAM,YAAY,gBAAgB,MAAM;AACxC,OAAI,SAAS,KAAK,UAAU;;AAE9B,MAAI,CAAC,KAAK;AACR,SAAM,sBAAsB,MAAM;AAClC,SAAO,SAAS,KAAK,IAAI;GACzB,MAAM,YAAY,gBAAgB,MAAM;AACxC,OAAI,SAAS,KAAK,UAAU;aACnB,IAAI,SAAS,WAAW,GAAG;GACpC,MAAM,YAAY,gBAAgB,MAAM;AACxC,OAAI,SAAS,KAAK,UAAU;;AAE9B,SAAO,gBAAgB,MAAM;AAC7B,MAAI,SAAS,KAAK,KAAK;AACvB,eAAa;AACb,sBAAoB;;CAGtB,MAAM,iBAAiB,UAA6B;AAClD,MAAI,CAAC,MACH,WAAU,MAAM;AAElB,MAAI,CAAC,KAAK;AACR,SAAM,sBAAsB,MAAM;AAClC,SAAO,SAAS,KAAK,IAAI;;AAE3B,MAAI,CAAC,MAAM;AACT,UAAO,gBAAgB,MAAM;AAC7B,OAAI,SAAS,KAAK,KAAK;;AAEzB,MAAI,CAAC,YAAY;AACf,gBAAa,mBAAmB,MAAM;AACtC,QAAK,SAAS,KAAK,WAAW;AAC9B,uBAAoB;;;CAIxB,MAAM,wBAAwB,WAA4B;AACxD,MAAI,YAAY,SACd,YAAW,SAAS,MAAM;AAE5B,MAAI,MAAM,SACR,MAAK,SAAS,MAAM;AAEtB,MAAI,KAAK,SACP,KAAI,SAAS,MAAM;AAErB,MAAI,OAAO,SACT,OAAM,SAAS,MAAM;AAEvB,MAAI,KAAK,SACP,KAAI,SAAS,MAAM;;CAIvB,MAAM,gBAAgB,QAAe;AACnC,UAAQ,IAAI,MAAZ;GACE,KAAK;AAEH,yBADe,kBAAkB,IAAI,SAAS,MAClB;AAC5B,qCAAiC;AACjC,gBAAY;AACZ;GAEF,KAAK;AACH,qBAAiB,IAAI,SAAS;AAC9B,kBAAc,IAAI,SAAS;AAE3B,QAAI,oBAAoB;AACtB,0BAAqB;AACrB;;AAGF,QAAI,CAAC,MACH,WAAU,IAAI,SAAS,MAAM;AAE/B,cAAU,IAAI,SAAS,IAAI;AAC3B;GAEF,KAAK;AACH,qBAAiB,IAAI,SAAS;AAC9B,kBAAc,IAAI,SAAS;AAC3B,QAAI,CAAC,MACH,WAAU,IAAI,SAAS,MAAM;AAE/B,mBAAe,IAAI,SAAS,IAAI;AAChC;GAEF,KAAK;AACH,qBAAiB,IAAI,SAAS;AAC9B,kBAAc,IAAI,SAAS;AAC3B,kBAAc,IAAI,SAAS,IAAI;AAC/B;GAEF,KAAK;AACH,qBAAiB,IAAI,SAAS;AAC9B,kBAAc,IAAI,SAAS;AAC3B,oBAAgB,IAAI,SAAS,MAAM;AAEnC,QAAI,KAAM,SAAS,WAAW,GAAG;KAC/B,MAAM,WAAW,mBAAmB,IAAI,SAAS,MAAM;AACvD,UAAM,SAAS,KAAK,SAAS;;AAE/B,iBAAa,mBAAmB,IAAI,SAAS,IAAI;AACjD,UAAM,SAAS,KAAK,WAAW;AAC/B,wBAAoB;AACpB;GAEF,KAAK,QAAQ;IACX,MAAM,MAAM,IAAI,SAAS;AAGzB,QAAI,CAAC,cAAc;AACjB,UAAK,WAAW;MACd,KAAK,IAAI,SAAS;MAClB,OAAO,IAAI,SAAS;MACrB;AACD,oBAAe;;AAIjB,QAAI,sBAAsB;AACxB,iBAAY,KAAK,IAAI,SAAS;AAC9B,sBAAiB,IAAI,SAAS;AAC9B,mBAAc,IAAI,SAAS;AAC3B;;AAIF,kBAAc,IAAI,SAAS,MAAM;AACjC,QAAI,WACF,YAAW,SAAS;AAEtB,yBAAqB,IAAI,SAAS,IAAI;AACtC,qBAAiB,IAAI,SAAS;AAC9B,kBAAc,IAAI,SAAS;AAC3B,wBAAoB;AACpB;;GAEF,QACE,OAAM,IAAI,MAAM,0BAA0B,IAAI,OAAO;;;CAO3D,MAAM,iBAAiB;AACrB,MAAI,kBAAkB,IACpB,sBAAqB,eAAe;AAEtC,mCAAiC;AACjC,MAAI,OAAO,CAAC,qBAAqB,CAAC,IAAI,KASpC,MAAK,SAAS,KAAK;AAIrB,MAAI,KAAK,YAAY,YACnB,MAAK,SAAS,MAAM;WACX,CAAC,KAAK,SACf,MAAK,WAAW;GACd,KAAK;IAAE,QAAQ;IAAG,MAAM;IAAG,QAAQ;IAAG;GACtC,OAAO;IAAE,QAAQ;IAAG,MAAM;IAAG,QAAQ;IAAG;GACzC;AAGH,SAAO;;;;;;;CAQT,SAAS,kCAAkC;AACzC,MAAI,CAAC,OAAO,IAAI,SAAS,WAAW,EAClC;EAEF,MAAM,YAAY,IAAI,SAAS,GAAG,GAAG;AACrC,MAAI,WAAW,SAAS,QACtB;EAEF,MAAM,YAAY;AAGlB,MAAI,YAAY,UAAU,IAAI,CAAC,qBAAqB,UAAU,CAC5D,KAAI,SAAS,KAAK;;AAItB,QAAO;EAAE;EAAU;EAAc;EAAM;;AAGzC,SAAgB,uBAAuB,QAA+B;CACpE,MAAM,OAAO,kBAAkB;AAC/B,MAAK,MAAM,OAAO,OAChB,MAAK,aAAa,IAAI;AAExB,QAAO,KAAK,UAAU;;;;ACxXxB,MAAM,oBAAoB;AAC1B,MAAM,kBAAkB;AACxB,MAAM,4BAA4B;AAClC,MAAM,0BAA0B;AAChC,MAAM,4BAA4B;AAClC,MAAM,0BAA0B;AAEhC,IAAa,iBAAb,MAAkE;CAChE,QAAgB;CAChB,IAAY;CACZ,OAAe;CACf,MAAc;CACd;CAGA,kBAA0B;CAC1B,mBAGI;CAEJ,MAAM,KAAoB;AACxB,OAAK,QAAQ,IAAI;AACjB,OAAK,SAAS,IAAI;AAClB,OAAK,IAAI;AACT,OAAK,OAAO;AACZ,OAAK,MAAM;AACX,OAAK,kBAAkB;AACvB,OAAK,mBAAmB;AAGxB,MAAI,KAAK,MAAM,WAAW,MAAM,EAAE;GAEhC,MAAM,MAAM,KAAK,OAAO,mBAAmB,gBAAgB;GAC3D,MAAM,OAAO,KAAK,OAChB,2BACA,wBACD;GACD,MAAM,OAAO,KAAK,OAChB,2BACA,wBACD;AACD,QAAK,mBAAmB;IACtB;KAAE,SAAS,IAAI;KAAQ,MAAM;KAAQ,OAAO;KAAK;IACjD;KAAE,SAAS;KAAG,MAAM;KAAe;IACnC;KAAE,SAAS,KAAK;KAAQ,MAAM;KAAQ,OAAO;KAAM;IACnD;KAAE,SAAS;KAAG,MAAM;KAAe;IACnC;KAAE,SAAS,KAAK;KAAQ,MAAM;KAAQ,OAAO;KAAM;IACpD;;;CAML,OAAqB;EACnB,MAAM,IAAI,KAAK;EACf,MAAM,IAAI,EAAE;AACZ,MAAI,KAAK,KAAK,KAAK,CAAC,KAAK,kBAAkB,OACzC,QAAO;EAGT,MAAM,QAAQ;GAAE,QAAQ,KAAK;GAAK,MAAM,KAAK;GAAM,QAAQ,KAAK;GAAG;AAGnE,MAAI,CAAC,KAAK,mBAAmB,KAAK,kBAAkB;GAClD,MAAM,OAAO,KAAK,iBAAiB,OAAO;AAC1C,OAAI,MAAM;AACR,QAAI,KAAK,SAAS,QAAQ;KACxB,MAAM,OAAO,KAAK,IAAI,KAAK,SAAS,IAAI,KAAK,EAAE;KAC/C,MAAM,MAAM,KAAK,MAAM,MAAM,GAAG,KAAK;AACrC,UAAK,aAAa,IAAI;AACtB,SAAI,KAAK,iBAAiB,WAAW,GAAG;AACtC,WAAK,mBAAmB;AACxB,WAAK,kBAAkB;;AAEzB,YAAO,KAAK,KAAK,QAAQ,KAAK,MAAM;;AAGtC,SAAK,SAAS,KAAK,IAAI,KAAK,SAAS,IAAI,KAAK,EAAE,CAAC;AACjD,QAAI,KAAK,iBAAiB,WAAW,GAAG;AACtC,UAAK,mBAAmB;AACxB,UAAK,kBAAkB;;AAEzB,WAAO,KAAK,KAAK,eAAe,KAAA,GAAW,MAAM;;;AAKrD,MAAI,KAAK,KAAK,EACZ,QAAO;EAET,MAAM,KAAK,EAAE,OAAO,KAAK,EAAE;AAE3B,MAAI,OAAO,KAAK,OAAO,SAAS;AAC9B,QAAK,SAAS,GAAG,KAAK;AACtB,UAAO,KAAK,KAAK,eAAe,KAAA,GAAW,MAAM;;AAEnD,MAAI,OAAO,KAAK,OAAO,OAAO;AAC5B,QAAK,SAAS,EAAE;AAChB,UAAO,KAAK,KAAK,eAAe,KAAA,GAAW,MAAM;;AAEnD,MAAI,OAAO,KAAK,OAAO,YAAY;AACjC,QAAK,SAAS,EAAE;AAChB,UAAO,KAAK,KAAK,oBAAoB,KAAA,GAAW,MAAM;;AAExD,MAAI,OAAO,KAAK,OAAO,WAAW;AAChC,QAAK,SAAS,EAAE;AAChB,UAAO,KAAK,KAAK,mBAAmB,KAAA,GAAW,MAAM;;AAEvD,MAAI,OAAO,KAAK,OAAO,cAAc;AACnC,QAAK,SAAS,EAAE;AAChB,UAAO,KAAK,KAAK,iBAAiB,KAAA,GAAW,MAAM;;EAIrD,IAAI,IAAI,KAAK;EACb,MAAM,MAAM,KAAK,OAAO,WACtB,MAAM,KAAK,OAAO,OAClB,MAAM,KAAK,OAAO,YAClB,MAAM,KAAK,OAAO,SAClB,MAAM,KAAK,OAAO;AAEpB,SAAO,IAAI,EAAE,QAAQ;GACnB,MAAM,IAAI,EAAE,OAAO,EAAE;AACrB,OAAI,MAAM,OAAO,MAAM,OAAO,MAAM,OAAO,MAAM,OAAO,MAAM,IAC5D;AAEF,QAAK;;EAGP,MAAM,MAAM,EAAE,MAAM,KAAK,GAAG,EAAE;AAC9B,OAAK,aAAa,IAAI;AACtB,SAAO,KAAK,KAAK,QAAQ,KAAK,MAAM;;CAItC,OAAe,OAAe,KAAqB;AACjD,SAAO,KAAK,MAAM,MAAM,OAAO,KAAK,IAAI,KAAK,KAAK,MAAM,OAAO,CAAC;;CAGlE,SAAiB,GAAW,YAAY,OAAO;AAC7C,OAAK,KAAK;AACV,MAAI,WAAW;AACb,QAAK,QAAQ;AACb,QAAK,MAAM;QAEX,MAAK,OAAO;;CAIhB,aAAqB,OAAe;EAClC,MAAM,QAAQ,MAAM,MAAM,KAAK,OAAO,QAAQ;AAC9C,MAAI,MAAM,SAAS,GAAG;AACpB,QAAK,QAAQ,MAAM,SAAS;AAC5B,QAAK,OAAO,MAAM,GAAG,GAAG,EAAE,UAAU,KAAK;QAEzC,MAAK,OAAO,MAAM;AAEpB,OAAK,KAAK,MAAM;;CAGlB,KACE,MACA,OACA,OACO;AAEP,SAAO;GAAE,UAAU;IAAE,KAAA;KADP,QAAQ,KAAK;KAAK,MAAM,KAAK;KAAM,QAAQ,KAAK;KACtC;IAAE;IAAO;GAAE;GAAM;GAAO;;CAIlD,CAAC,OAAO,YAA6B;AACnC,SAAO,EACL,YAAY;GACV,MAAM,IAAI,KAAK,MAAM;AACrB,OAAI,CAAC,EACH,QAAO;IAAE,MAAM;IAAM,OAAO,KAAA;IAA+B;AAE7D,UAAO;IAAE,MAAM;IAAO,OAAO;IAAG;KAEnC;;;;;AChLL,UAAU,qBAAqB,GAA+B;AAC5D,MAAK,IAAI,MAAM,EAAE,MAAM,EAAE,KAAK,MAAM,EAAE,MAAM,CAC1C,OAAM;;AAIV,SAAgB,WACd,OACA,OAAqB,EAAE,EACvB,UACM;CAKN,IAAI,MAAqB;EACvB,YAAY;GACV,GAAG;GACH,GAAI,UAAU,cAAc,EAAE;GAC/B;EACD;EACD;AAED,OAAM,iBAAiB,KAAK,KAAK,cAAc,qBAAqB;CAEpE,MAAM,YAAY,IAAI,gBAAgB;AAEtC,WAAU,MAAM,IAAI;AAEpB,QAAO,uBAAuB,qBAAqB,UAAU,CAAC;;AAGhE,MAAM,cAAqD,SAAS,YAElE,UAAU,EAAE,EACZ;AACA,MAAK,UAAU,aAAqB;AAIlC,SAAO,WAAW,UAAU,SAFX,KAAK,KAAK,WAEkB,CAAC"}

package/dist/types.d.ts

@@ -1,5 +1,15 @@
-import type { Delimiters } from "@glion/ast";
+import type { Delimiters, Root } from "@glion/ast";
+import type { Processor } from "unified";
 import type { Position } from "unist";
+/**
+ * A unified `Processor` that parses HL7v2 messages into `Root` trees.
+ *
+ * Only `ParseTree` is constrained — head and tail admit `undefined` so that
+ * any processor built with `unified().use(hl7v2Parser)` is assignable, from
+ * the bare parser (whose transform head/tail are `undefined`) through full
+ * transformer/compiler pipelines like `@glion/hl7v2`'s `parseHL7v2`.
+ */
+export type Hl7v2Processor = Processor<Root, Root | undefined, Root | undefined>;
 export type PreprocessorStep = (ctx: ParserContext) => ParserContext;
 export interface ParseOptions {
     /**

package/dist/types.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../src/types.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,YAAY,CAAC;AAC7C,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,OAAO,CAAC;AAItC,MAAM,MAAM,gBAAgB,GAAG,CAAC,GAAG,EAAE,aAAa,KAAK,aAAa,CAAC;AAErE,MAAM,WAAW,YAAY;IAC3B;;OAEG;IACH,UAAU,CAAC,EAAE,gBAAgB,EAAE,CAAC;CACjC;AAED,MAAM,WAAW,aAAa;IAC5B,KAAK,EAAE,MAAM,CAAC;IACd,UAAU,EAAE,UAAU,CAAC;IACvB,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;CACpC;AAGD,MAAM,WAAW,KAAK;IACpB,IAAI,EAAE,SAAS,CAAC;IAChB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,QAAQ,EAAE,QAAQ,CAAC;CACpB;AAGD,MAAM,MAAM,SAAS,GACjB,aAAa,GACb,aAAa,GACb,kBAAkB,GAClB,iBAAiB,GACjB,eAAe,GACf,MAAM,CAAC;AAEX,MAAM,WAAW,SAAS;IACxB,KAAK,CAAC,GAAG,EAAE,aAAa,GAAG,IAAI,CAAC;IAChC,IAAI,IAAI,KAAK,GAAG,IAAI,CAAC;CACtB"}
+{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../src/types.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,IAAI,EAAE,MAAM,YAAY,CAAC;AACnD,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,SAAS,CAAC;AACzC,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,OAAO,CAAC;AAEtC;;;;;;;GAOG;AACH,MAAM,MAAM,cAAc,GAAG,SAAS,CACpC,IAAI,EACJ,IAAI,GAAG,SAAS,EAChB,IAAI,GAAG,SAAS,CACjB,CAAC;AAIF,MAAM,MAAM,gBAAgB,GAAG,CAAC,GAAG,EAAE,aAAa,KAAK,aAAa,CAAC;AAErE,MAAM,WAAW,YAAY;IAC3B;;OAEG;IACH,UAAU,CAAC,EAAE,gBAAgB,EAAE,CAAC;CACjC;AAED,MAAM,WAAW,aAAa;IAC5B,KAAK,EAAE,MAAM,CAAC;IACd,UAAU,EAAE,UAAU,CAAC;IACvB,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;CACpC;AAGD,MAAM,WAAW,KAAK;IACpB,IAAI,EAAE,SAAS,CAAC;IAChB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,QAAQ,EAAE,QAAQ,CAAC;CACpB;AAGD,MAAM,MAAM,SAAS,GACjB,aAAa,GACb,aAAa,GACb,kBAAkB,GAClB,iBAAiB,GACjB,eAAe,GACf,MAAM,CAAC;AAEX,MAAM,WAAW,SAAS;IACxB,KAAK,CAAC,GAAG,EAAE,aAAa,GAAG,IAAI,CAAC;IAChC,IAAI,IAAI,KAAK,GAAG,IAAI,CAAC;CACtB"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@glion/parser",
-  "version": "0.16.0",
+  "version": "0.20.0",
   "description": "hl7v2 plugin to parse hl7v2 messages",
   "keywords": [
     "health",
@@ -33,35 +33,36 @@
     "access": "public"
   },
   "dependencies": {
+    "@types/unist": "3.0.3",
     "unified": "11.0.5",
-    "@glion/utils": "0.16.0"
+    "@glion/ast": "0.20.0",
+    "@glion/config": "0.20.0",
+    "@glion/utils": "0.20.0"
   },
   "devDependencies": {
+    "@stryker-mutator/core": "10.0.0",
+    "@stryker-mutator/vitest-runner": "10.0.0",
     "@types/node": "25.6.0",
-    "@types/unist": "^3.0.3",
-    "@vitest/coverage-v8": "4.1.5",
+    "@vitest/coverage-v8": "4.1.11",
     "publint": "0.3.18",
     "tsdown": "0.21.10",
     "typescript": "6.0.3",
-    "unist": "^0.0.1",
-    "vitest": "4.1.5",
-    "@glion/ast": "0.16.0",
-    "@glion/check-readme": "0.16.0",
-    "@glion/config": "0.16.0",
-    "@glion/testing": "0.16.0",
-    "@glion/tsconfig": "0.16.0"
+    "vitest": "4.1.11",
+    "@glion/check-readme": "0.18.0",
+    "@glion/testing": "0.18.0",
+    "@glion/tsconfig": "0.18.0"
   },
   "engines": {
-    "node": ">=20"
+    "node": ">=22"
   },
   "scripts": {
-    "bench": "vitest bench --run",
     "build": "tsdown && tsc --emitDeclarationOnly",
-    "check-types": "tsc --noEmit",
+    "check-types": "tsc --noEmit && tsc --noEmit -p tests/tsconfig.json",
     "check:readme": "glion-check-readme",
+    "mutate": "stryker run",
+    "publint": "publint --strict",
     "test": "vitest run",
     "test:coverage": "vitest run --coverage",
-    "test:watch": "vitest",
-    "publint": "publint --strict"
+    "test:watch": "vitest"
   }
 }