@cosyte/cli 0.0.6 → 0.1.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -3,7 +3,8 @@ import { parseArgs } from 'util';
3
3
  import { Server } from '@modelcontextprotocol/sdk/server/index.js';
4
4
  import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio.js';
5
5
  import { ListToolsRequestSchema, CallToolRequestSchema } from '@modelcontextprotocol/sdk/types.js';
6
- import { readFile } from 'fs/promises';
6
+ import { open } from 'fs/promises';
7
+ import { randomBytes } from 'crypto';
7
8
 
8
9
  var __defProp = Object.defineProperty;
9
10
  var __getOwnPropNames = Object.getOwnPropertyNames;
@@ -27,21 +28,29 @@ var init_exit_codes = __esm({
27
28
  EXIT = {
28
29
  /** Success: the operation completed; `validate` found the input **valid**. */
29
30
  OK: 0,
30
- /** Operation-level failure: `validate` found the input **invalid** (parseable but non-conformant):
31
- * a real, expected CI signal that the message is bad and the tool worked. Never emitted for
32
- * unparseable input (that is `DATAERR`). */
31
+ /** Operation-level failure: `validate` found the input **invalid** (parseable but non-conformant),
32
+ * or `redact` could not de-identify every locus: a real, expected CI signal that the tool worked
33
+ * and the operation did not complete. Never emitted for unparseable input (that is `DATAERR`). */
33
34
  INVALID: 1,
34
- /** Usage error: unknown command, bad flag, missing argument (`EX_USAGE`). */
35
+ /** Usage error: unknown command, bad flag, missing argument. A value specific to this CLI, not a
36
+ * `sysexits.h` constant. */
35
37
  USAGE: 2,
36
38
  /** Data error: input could not be parsed or its format could not be detected (`EX_DATAERR`). */
37
39
  DATAERR: 65,
38
40
  /** No input: the named file does not exist or is unreadable (`EX_NOINPUT`). */
39
41
  NOINPUT: 66,
40
- /** Unavailable: a required capability is not yet wired (e.g. `redact` before `@cosyte/deid`), a
41
- * distinct non-zero signal that is never a fake success (`EX_UNAVAILABLE`). */
42
+ /** Unavailable: a required capability is not wired for this input (e.g. `redact` on a format
43
+ * `@cosyte/deid` has no adapter for, or with that optional library absent), a distinct non-zero
44
+ * signal that is never a fake success (`EX_UNAVAILABLE`). */
42
45
  UNAVAILABLE: 69,
43
46
  /** Internal error: an unexpected exception (a bug), distinct from a handled bad input (`EX_SOFTWARE`). */
44
- SOFTWARE: 70
47
+ SOFTWARE: 70,
48
+ /** Output error: the stream the CLI was writing to closed before the output had been delivered, so
49
+ * the consumer holds nothing or holds only part of the answer. The shape a downstream consumer
50
+ * that exits early produces (a pager quit, a `head` that has seen enough). A handled condition the
51
+ * consumer owns, never a defect in the CLI, so it is deliberately distinct from `SOFTWARE`
52
+ * (`EX_IOERR`). */
53
+ IOERR: 74
45
54
  };
46
55
  }
47
56
  });
@@ -86,12 +95,27 @@ var init_diagnostics = __esm({
86
95
  CLI_PARSER_UNAVAILABLE: "CLI_PARSER_UNAVAILABLE",
87
96
  /** The wrapped parser rejected the input. Positional context only, never the offending bytes. Exit `65`. */
88
97
  CLI_PARSE_FAILED: "CLI_PARSE_FAILED",
98
+ /** The input is larger than the documented size limit the CLI reads in one invocation. Raised
99
+ * against the running byte count while the input is read, so the refusal always arrives before an
100
+ * allocation ceiling could. Names the limit, never the input. Exit `65`, never `70`. */
101
+ CLI_INPUT_TOO_LARGE: "CLI_INPUT_TOO_LARGE",
102
+ /** The output stream could not be written: a consumer that closes the pipe ends the invocation
103
+ * here rather than as an unhandled platform error, whether it went away part way through a record
104
+ * stream or before a single-result write was delivered. Names no input, and never the platform's
105
+ * own error text. Exit `74`. */
106
+ CLI_OUTPUT_WRITE_FAILED: "CLI_OUTPUT_WRITE_FAILED",
89
107
  /** The BYO ConceptMap supplied to `map-codes` is not valid JSON or not a loadable FHIR ConceptMap.
90
108
  * Names the stable terminology-loader code, never the map's bytes. Exit `65`. */
91
109
  CLI_MAP_INVALID: "CLI_MAP_INVALID",
92
- /** A command whose ground-layer library is not yet wired (e.g. `redact` before `@cosyte/deid` is
93
- * wired). Never a fake success: a distinct, value-free "unavailable" signal. Exit `69`. */
110
+ /** A command whose ground-layer library does not cover the resolved format (e.g. `redact` on a
111
+ * format `@cosyte/deid` ships no adapter for). Never a fake success: a distinct, value-free
112
+ * "unavailable" signal. Exit `69`. */
94
113
  CLI_NOT_IMPLEMENTED: "CLI_NOT_IMPLEMENTED",
114
+ /** The delegated de-identifier reported that it could not handle every element of the input (a
115
+ * fail-closed `blocked` locus, or its own fatal). The run is **not** a de-identified copy, so no
116
+ * output is emitted: an operation-level failure, never a bug. Names loci and codes, never a
117
+ * value. Exit `1`. */
118
+ CLI_DEID_INCOMPLETE: "CLI_DEID_INCOMPLETE",
95
119
  /** An unexpected internal error (a bug). Exit `70`. */
96
120
  CLI_INTERNAL: "CLI_INTERNAL"
97
121
  };
@@ -115,6 +139,28 @@ var init_diagnostics = __esm({
115
139
  }
116
140
  });
117
141
 
142
+ // src/core/limits.ts
143
+ function describeByteLimit(limit = MAX_INPUT_BYTES) {
144
+ const mib = limit / BYTES_PER_MIB;
145
+ return Number.isInteger(mib) ? `${String(limit)} bytes (${String(mib)} MiB)` : `${String(limit)} bytes`;
146
+ }
147
+ function inputTooLargeError(limit = MAX_INPUT_BYTES) {
148
+ return new CliError(
149
+ CLI_CODES.CLI_INPUT_TOO_LARGE,
150
+ EXIT.DATAERR,
151
+ `input is larger than the ${describeByteLimit(limit)} this command reads in one invocation; split the input and re-run`
152
+ );
153
+ }
154
+ var BYTES_PER_MIB, MAX_INPUT_BYTES;
155
+ var init_limits = __esm({
156
+ "src/core/limits.ts"() {
157
+ init_diagnostics();
158
+ init_exit_codes();
159
+ BYTES_PER_MIB = 1024 * 1024;
160
+ MAX_INPUT_BYTES = 64 * BYTES_PER_MIB;
161
+ }
162
+ });
163
+
118
164
  // src/core/format.ts
119
165
  function leadingText(bytes) {
120
166
  const slice = bytes.subarray(0, SNIFF_BYTES);
@@ -190,12 +236,13 @@ function detectionError(detected) {
190
236
  function asCosyteFormat(value) {
191
237
  return KNOWN_FORMATS.includes(value) ? value : null;
192
238
  }
193
- var SNIFF_BYTES, SIGNATURES, DETECTABLE_FORMATS, KNOWN_FORMATS;
239
+ var DETECT_PREFIX_BYTES, SNIFF_BYTES, SIGNATURES, DETECTABLE_FORMATS, KNOWN_FORMATS;
194
240
  var init_format = __esm({
195
241
  "src/core/format.ts"() {
196
242
  init_diagnostics();
197
243
  init_exit_codes();
198
- SNIFF_BYTES = 512;
244
+ DETECT_PREFIX_BYTES = 512;
245
+ SNIFF_BYTES = DETECT_PREFIX_BYTES;
199
246
  SIGNATURES = [
200
247
  { format: "hl7", match: (prefix) => looksLikeHl7(prefix) },
201
248
  { format: "mllp", match: (prefix, bytes) => looksLikeMllp(prefix, bytes) },
@@ -561,11 +608,7 @@ async function validateFormat(format, bytes) {
561
608
  }
562
609
  async function deframeMllp(bytes) {
563
610
  if (bytes.lastIndexOf(11) > bytes.lastIndexOf(28)) {
564
- throw new CliError(
565
- CLI_CODES.CLI_PARSE_FAILED,
566
- EXIT.DATAERR,
567
- "truncated MLLP stream: an unterminated frame at the end of input (no closing FS/CR)"
568
- );
611
+ throw truncatedMllpError();
569
612
  }
570
613
  const { FrameReader } = await loadOptional("mllp", () => import('@cosyte/mllp'));
571
614
  const payloads = [];
@@ -577,6 +620,13 @@ async function deframeMllp(bytes) {
577
620
  reader.push(Buffer.from(bytes));
578
621
  return { payloads, warningCount };
579
622
  }
623
+ function truncatedMllpError() {
624
+ return new CliError(
625
+ CLI_CODES.CLI_PARSE_FAILED,
626
+ EXIT.DATAERR,
627
+ "truncated MLLP stream: an unterminated frame at the end of input (no closing FS/CR)"
628
+ );
629
+ }
580
630
  function decode(bytes) {
581
631
  return new TextDecoder("utf-8", { fatal: false }).decode(bytes);
582
632
  }
@@ -612,6 +662,91 @@ var init_parsers = __esm({
612
662
  }
613
663
  });
614
664
 
665
+ // src/core/records.ts
666
+ function asBuffer(chunk) {
667
+ return Buffer.isBuffer(chunk) ? chunk : Buffer.from(chunk.buffer, chunk.byteOffset, chunk.length);
668
+ }
669
+ async function* oneChunk(bytes) {
670
+ await Promise.resolve();
671
+ if (bytes.length > 0) yield bytes;
672
+ }
673
+ async function* withinLimit(chunks, limit = MAX_INPUT_BYTES) {
674
+ let total = 0;
675
+ for await (const chunk of chunks) {
676
+ total += chunk.length;
677
+ if (total > limit) throw inputTooLargeError(limit);
678
+ yield chunk;
679
+ }
680
+ }
681
+ async function collectChunks(chunks) {
682
+ const parts = [];
683
+ for await (const chunk of chunks) parts.push(asBuffer(chunk));
684
+ return Buffer.concat(parts);
685
+ }
686
+ function isBlank(line) {
687
+ return new TextDecoder("utf-8", { fatal: false }).decode(line).trim().length === 0;
688
+ }
689
+ function stripCarriageReturn(line) {
690
+ return line.length > 0 && line[line.length - 1] === CARRIAGE_RETURN ? line.subarray(0, line.length - 1) : line;
691
+ }
692
+ async function* ndjsonRecords(chunks) {
693
+ let carry = Buffer.alloc(0);
694
+ for await (const chunk of chunks) {
695
+ const buf = carry.length > 0 ? Buffer.concat([carry, asBuffer(chunk)]) : asBuffer(chunk);
696
+ let start = 0;
697
+ for (; ; ) {
698
+ const end = buf.indexOf(LINE_FEED, start);
699
+ if (end < 0) break;
700
+ const line = stripCarriageReturn(buf.subarray(start, end));
701
+ if (!isBlank(line)) yield line;
702
+ start = end + 1;
703
+ }
704
+ carry = start === 0 ? asBuffer(buf) : Buffer.from(buf.subarray(start));
705
+ }
706
+ const last = stripCarriageReturn(carry);
707
+ if (!isBlank(last)) yield last;
708
+ }
709
+ async function* mllpFrames(chunks) {
710
+ const { FrameReader } = await loadOptional("mllp", () => import('@cosyte/mllp'));
711
+ const ready = [];
712
+ const reader = new FrameReader({
713
+ onFrame: (payload) => ready.push(Buffer.from(payload)),
714
+ onWarning: () => void 0,
715
+ // The CLI's own documented input limit is the binding one, so the reader's smaller default frame
716
+ // ceiling is raised to it: an over-limit MLLP stream is refused by the limit the help output and
717
+ // the command reference state, not by an undocumented one underneath it.
718
+ maxFrameSizeBytes: MAX_INPUT_BYTES
719
+ });
720
+ let lastStart = -1;
721
+ let lastEnd = -1;
722
+ let offset = 0;
723
+ for await (const chunk of chunks) {
724
+ const buf = asBuffer(chunk);
725
+ const start = buf.lastIndexOf(VERTICAL_TAB);
726
+ if (start >= 0) lastStart = offset + start;
727
+ const end = buf.lastIndexOf(FILE_SEPARATOR);
728
+ if (end >= 0) lastEnd = offset + end;
729
+ offset += buf.length;
730
+ reader.push(buf);
731
+ while (ready.length > 0) {
732
+ const frame = ready.shift();
733
+ if (frame !== void 0) yield frame;
734
+ }
735
+ }
736
+ if (lastStart > lastEnd) throw truncatedMllpError();
737
+ }
738
+ var LINE_FEED, CARRIAGE_RETURN, VERTICAL_TAB, FILE_SEPARATOR;
739
+ var init_records = __esm({
740
+ "src/core/records.ts"() {
741
+ init_limits();
742
+ init_parsers();
743
+ LINE_FEED = 10;
744
+ CARRIAGE_RETURN = 13;
745
+ VERTICAL_TAB = 11;
746
+ FILE_SEPARATOR = 28;
747
+ }
748
+ });
749
+
615
750
  // src/core/input.ts
616
751
  async function resolveInput(source, formatOverride, deps2, op) {
617
752
  if (source === void 0) {
@@ -633,34 +768,91 @@ async function resolveInput(source, formatOverride, deps2, op) {
633
768
  if (bytes.length === 0) {
634
769
  return fail(new CliError(CLI_CODES.CLI_EMPTY_INPUT, EXIT.DATAERR, "input is empty"));
635
770
  }
771
+ const format = resolveFormat(bytes, formatOverride, op);
772
+ if (format instanceof CliError) return fail(format);
773
+ return { ok: true, input: { format, bytes } };
774
+ }
775
+ async function resolveInputStream(source, formatOverride, deps2, op, limit = MAX_INPUT_BYTES) {
776
+ if (source === void 0) {
777
+ return fail(
778
+ new CliError(
779
+ CLI_CODES.CLI_USAGE,
780
+ EXIT.USAGE,
781
+ "missing <file> argument; pass a path or `-` to read stdin"
782
+ )
783
+ );
784
+ }
785
+ const iterator = withinLimit(openChunks(source, deps2), limit)[Symbol.asyncIterator]();
786
+ const need = formatOverride === void 0 ? DETECT_PREFIX_BYTES : 1;
787
+ const head = [];
788
+ let headBytes = 0;
789
+ try {
790
+ while (headBytes < need) {
791
+ const step = await iterator.next();
792
+ if (step.done === true) break;
793
+ head.push(step.value);
794
+ headBytes += step.value.length;
795
+ }
796
+ } catch (e) {
797
+ if (e instanceof CliError) return fail(e);
798
+ throw e;
799
+ }
800
+ if (headBytes === 0) {
801
+ return fail(new CliError(CLI_CODES.CLI_EMPTY_INPUT, EXIT.DATAERR, "input is empty"));
802
+ }
803
+ const prefix = Buffer.concat(
804
+ head.map((c) => Buffer.from(c.buffer, c.byteOffset, c.length)),
805
+ Math.min(headBytes, DETECT_PREFIX_BYTES)
806
+ );
807
+ const format = resolveFormat(prefix, formatOverride, op);
808
+ if (format instanceof CliError) {
809
+ await iterator.return(void 0);
810
+ return fail(format);
811
+ }
812
+ return { ok: true, input: { format, chunks: rejoin(head, iterator) } };
813
+ }
814
+ function openChunks(source, deps2) {
815
+ if (source === "-") {
816
+ return deps2.openStdin === void 0 ? readerAsChunks(() => deps2.readStdin()) : deps2.openStdin();
817
+ }
818
+ return deps2.openFile === void 0 ? readerAsChunks(() => deps2.readFile(source)) : deps2.openFile(source);
819
+ }
820
+ async function* readerAsChunks(read) {
821
+ yield* oneChunk(await read());
822
+ }
823
+ async function* rejoin(head, rest) {
824
+ for (const chunk of head) yield chunk;
825
+ for (; ; ) {
826
+ const step = await rest.next();
827
+ if (step.done === true) return;
828
+ yield step.value;
829
+ }
830
+ }
831
+ function resolveFormat(prefix, formatOverride, op) {
636
832
  let format;
637
833
  if (formatOverride !== void 0) {
638
834
  const narrowed = asCosyteFormat(formatOverride);
639
835
  if (narrowed === null) {
640
- return fail(
641
- new CliError(
642
- CLI_CODES.CLI_USAGE,
643
- EXIT.USAGE,
644
- "unknown --format value; expected one of hl7, fhir, dicom, x12, ccda, ncpdp, astm, mllp"
645
- )
836
+ return new CliError(
837
+ CLI_CODES.CLI_USAGE,
838
+ EXIT.USAGE,
839
+ "unknown --format value; expected one of hl7, fhir, dicom, x12, ccda, ncpdp, astm, mllp"
646
840
  );
647
841
  }
648
842
  format = narrowed;
649
843
  } else {
650
- const detected = detectFormat(bytes);
651
- if (detected.format === null) return fail(detectionError(detected));
844
+ const detected = detectFormat(prefix);
845
+ if (detected.format === null) return detectionError(detected);
652
846
  format = detected.format;
653
847
  }
654
- if (!supportsOp(format, op)) {
655
- return fail(
656
- new CliError(
657
- CLI_CODES.CLI_FORMAT_UNSUPPORTED,
658
- EXIT.DATAERR,
659
- `format '${format}' does not support \`${op}\` in this CLI build (${op} supports: ${formatsSupporting(op).join(", ")})`
660
- )
848
+ if (op !== null && !supportsOp(format, op)) {
849
+ return new CliError(
850
+ CLI_CODES.CLI_FORMAT_UNSUPPORTED,
851
+ EXIT.DATAERR,
852
+ `format '${format}' does not support \`${op}\` in this CLI build (${op} supports: ${formatsSupporting(op).join(", ")})`
661
853
  );
662
854
  }
663
- return { ok: true, input: { format, bytes } };
855
+ return format;
664
856
  }
665
857
  function fail(e) {
666
858
  return { ok: false, result: errorResult(e) };
@@ -670,7 +862,9 @@ var init_input = __esm({
670
862
  init_format();
671
863
  init_diagnostics();
672
864
  init_exit_codes();
865
+ init_limits();
673
866
  init_parsers();
867
+ init_records();
674
868
  }
675
869
  });
676
870
 
@@ -972,11 +1166,24 @@ async function parseCommand(args, deps2, posture = VALUE_FREE) {
972
1166
  )
973
1167
  );
974
1168
  }
975
- const resolved = await resolveInput(positionals[0], values.format, deps2, "parse");
1169
+ const resolved = await resolveInputStream(positionals[0], values.format, deps2, "parse");
976
1170
  if (!resolved.ok) return resolved.result;
977
- const { format, bytes } = resolved.input;
1171
+ const { format, chunks } = resolved.input;
978
1172
  if (format === "mllp" || values.ndjson === true) {
979
- return await parseMulti(format, bytes, values.ndjson === true, values.quiet === true);
1173
+ return await parseMulti(
1174
+ format,
1175
+ chunks,
1176
+ values.ndjson === true,
1177
+ values.quiet === true,
1178
+ deps2.writeStdout
1179
+ );
1180
+ }
1181
+ let bytes;
1182
+ try {
1183
+ bytes = await collectChunks(chunks);
1184
+ } catch (e) {
1185
+ if (e instanceof CliError) return errorResult(e);
1186
+ throw e;
980
1187
  }
981
1188
  return await parseSingle(format, bytes, values.json === true, values.quiet === true, posture);
982
1189
  }
@@ -995,23 +1202,57 @@ async function parseSingle(format, bytes, json, quiet, posture) {
995
1202
  ` : "";
996
1203
  return { stdout, stderr, exit: EXIT.OK };
997
1204
  }
998
- async function parseMulti(format, bytes, ndjson, quiet) {
999
- let records;
1000
- let recordFormat;
1205
+ async function parseMulti(format, chunks, ndjson, quiet, writeStdout) {
1206
+ const recordFormat = format === "mllp" ? "hl7" : format;
1207
+ const records = format === "mllp" ? mllpFrames(chunks) : ndjsonRecords(chunks);
1208
+ const held = [];
1209
+ const emit2 = (line) => {
1210
+ if (writeStdout === void 0) {
1211
+ held.push(line);
1212
+ return;
1213
+ }
1214
+ try {
1215
+ writeStdout(line);
1216
+ } catch {
1217
+ throw new CliError(
1218
+ CLI_CODES.CLI_OUTPUT_WRITE_FAILED,
1219
+ EXIT.IOERR,
1220
+ "could not write to the output stream; it closed before the record stream finished"
1221
+ );
1222
+ }
1223
+ };
1224
+ const written = () => held.join("");
1225
+ let total = 0;
1226
+ let failed = 0;
1227
+ let warnings = 0;
1001
1228
  try {
1002
- if (format === "mllp") {
1003
- const { payloads } = await deframeMllp(bytes);
1004
- records = payloads;
1005
- recordFormat = "hl7";
1006
- } else {
1007
- records = splitLines(bytes);
1008
- recordFormat = format;
1229
+ for await (const record of records) {
1230
+ let line;
1231
+ try {
1232
+ const { model, warnings: ws } = await parseFormat(recordFormat, record);
1233
+ warnings += ws.length;
1234
+ line = { record: total, format: recordFormat, model, warnings: ws };
1235
+ } catch (e) {
1236
+ if (e instanceof CliError) throw e;
1237
+ failed += 1;
1238
+ line = {
1239
+ record: total,
1240
+ format: recordFormat,
1241
+ error: extractStableCode(e) ?? "CLI_PARSE_FAILED"
1242
+ };
1243
+ }
1244
+ total += 1;
1245
+ emit2(JSON.stringify(line) + "\n");
1009
1246
  }
1010
1247
  } catch (e) {
1011
- if (e instanceof CliError) return errorResult(e);
1012
- return parseFailureResult(format, bytes, VALUE_FREE, e);
1248
+ if (e instanceof CliError) {
1249
+ return { stdout: written(), stderr: `${formatDiagnostic(e)}
1250
+ `, exit: e.exit };
1251
+ }
1252
+ const rejected = parseFailureResult(format, new Uint8Array(), VALUE_FREE, e);
1253
+ return { stdout: written(), stderr: rejected.stderr, exit: rejected.exit };
1013
1254
  }
1014
- if (records.length === 0) {
1255
+ if (total === 0) {
1015
1256
  return errorResult(
1016
1257
  new CliError(
1017
1258
  CLI_CODES.CLI_PARSE_FAILED,
@@ -1020,34 +1261,9 @@ async function parseMulti(format, bytes, ndjson, quiet) {
1020
1261
  )
1021
1262
  );
1022
1263
  }
1023
- const lines = [];
1024
- let failed = 0;
1025
- let warnings = 0;
1026
- for (let i = 0; i < records.length; i += 1) {
1027
- const rec = records[i];
1028
- try {
1029
- const { model, warnings: ws } = await parseFormat(recordFormat, rec);
1030
- warnings += ws.length;
1031
- lines.push({ record: i, format: recordFormat, model, warnings: ws });
1032
- } catch (e) {
1033
- if (e instanceof CliError) return errorResult(e);
1034
- failed += 1;
1035
- lines.push({
1036
- record: i,
1037
- format: recordFormat,
1038
- error: extractStableCode(e) ?? "CLI_PARSE_FAILED"
1039
- });
1040
- }
1041
- }
1042
- const stdout = lines.map((l) => JSON.stringify(l)).join("\n") + "\n";
1043
1264
  const exit = failed > 0 ? EXIT.DATAERR : EXIT.OK;
1044
- const stderr = quiet ? "" : `cosyte: parsed ${String(records.length)} ${recordFormat} record(s) (${String(warnings)} warning(s), ${String(failed)} failed)` + (ndjson ? " [ndjson]" : format === "mllp" ? " [mllp]" : "") + "\n";
1045
- return { stdout, stderr, exit };
1046
- }
1047
- function splitLines(bytes) {
1048
- const text = new TextDecoder("utf-8", { fatal: false }).decode(bytes);
1049
- const enc = new TextEncoder();
1050
- return text.split(/\r?\n/).filter((line) => line.trim().length > 0).map((line) => enc.encode(line));
1265
+ const stderr = quiet ? "" : `cosyte: parsed ${String(total)} ${recordFormat} record(s) (${String(warnings)} warning(s), ${String(failed)} failed)` + (ndjson ? " [ndjson]" : format === "mllp" ? " [mllp]" : "") + "\n";
1266
+ return { stdout: written(), stderr, exit };
1051
1267
  }
1052
1268
  var PARSE_OPTIONS;
1053
1269
  var init_parse = __esm({
@@ -1057,6 +1273,7 @@ var init_parse = __esm({
1057
1273
  init_input();
1058
1274
  init_parsers();
1059
1275
  init_phi();
1276
+ init_records();
1060
1277
  init_wrap();
1061
1278
  PARSE_OPTIONS = {
1062
1279
  format: { type: "string" },
@@ -1149,11 +1366,19 @@ var init_validate = __esm({
1149
1366
  var VERSION;
1150
1367
  var init_version = __esm({
1151
1368
  "src/core/version.ts"() {
1152
- VERSION = "0.0.6";
1369
+ VERSION = "0.1.1";
1153
1370
  }
1154
1371
  });
1155
1372
 
1156
1373
  // src/mcp/tools.ts
1374
+ function outputSchema(data) {
1375
+ return {
1376
+ type: "object",
1377
+ properties: { ...OUTCOME_PROPS, data },
1378
+ required: [...OUTCOME_REQUIRED],
1379
+ additionalProperties: false
1380
+ };
1381
+ }
1157
1382
  function inlineDeps(content) {
1158
1383
  const bytes = new TextEncoder().encode(content);
1159
1384
  const noFile = new CliError(
@@ -1172,37 +1397,65 @@ function stringArg(args, key) {
1172
1397
  const v = args[key];
1173
1398
  return typeof v === "string" ? v : null;
1174
1399
  }
1175
- function toToolResult(result) {
1176
- const hardError = result.stdout === "";
1177
- const text = hardError ? result.stderr.trim() : result.stdout.trim();
1400
+ function structuredResult(structuredContent) {
1178
1401
  return {
1179
- content: [
1180
- { type: "text", text: text.length > 0 ? text : `cosyte: exit ${String(result.exit)}` }
1181
- ],
1182
- isError: hardError,
1183
- structuredContent: { exit: result.exit, ok: !hardError }
1402
+ content: [{ type: "text", text: JSON.stringify(structuredContent) }],
1403
+ isError: structuredContent.status === "failed",
1404
+ structuredContent
1184
1405
  };
1185
1406
  }
1186
- function usageError(message) {
1187
- return {
1188
- content: [{ type: "text", text: `cosyte: ${CLI_CODES.CLI_USAGE}: ${message}` }],
1189
- isError: true,
1190
- structuredContent: { exit: EXIT.USAGE, ok: false }
1191
- };
1407
+ function isRecordLine(value) {
1408
+ return typeof value === "object" && value !== null && "record" in value;
1409
+ }
1410
+ function parsePayload(stdout) {
1411
+ const values = [];
1412
+ for (const line of stdout.split("\n")) {
1413
+ if (line.trim().length === 0) continue;
1414
+ try {
1415
+ values.push(JSON.parse(line));
1416
+ } catch {
1417
+ return void 0;
1418
+ }
1419
+ }
1420
+ const only = values.length === 1 ? values[0] : void 0;
1421
+ return only !== void 0 && !isRecordLine(only) ? only : { records: values };
1422
+ }
1423
+ function diagnosticCode(stderr) {
1424
+ return Object.values(CLI_CODES).find((code) => stderr.includes(`cosyte: ${code}: `));
1425
+ }
1426
+ function toToolResult(result) {
1427
+ if (result.stdout === "") {
1428
+ const code = diagnosticCode(result.stderr);
1429
+ return structuredResult({
1430
+ ok: false,
1431
+ status: "failed",
1432
+ exit: result.exit,
1433
+ ...code !== void 0 ? { code } : {}
1434
+ });
1435
+ }
1436
+ const data = parsePayload(result.stdout);
1437
+ return structuredResult({
1438
+ ok: true,
1439
+ status: result.exit === EXIT.OK ? "success" : "verdict",
1440
+ exit: result.exit,
1441
+ ...data !== void 0 ? { data } : {}
1442
+ });
1443
+ }
1444
+ function usageError() {
1445
+ return structuredResult({
1446
+ ok: false,
1447
+ status: "failed",
1448
+ exit: EXIT.USAGE,
1449
+ code: CLI_CODES.CLI_USAGE
1450
+ });
1192
1451
  }
1193
1452
  function internalError(e) {
1194
1453
  const err = toCliError(e);
1195
- return {
1196
- content: [{ type: "text", text: `cosyte: ${err.code}: ${err.message}` }],
1197
- isError: true,
1198
- structuredContent: { exit: err.exit, ok: false }
1199
- };
1454
+ return structuredResult({ ok: false, status: "failed", exit: err.exit, code: err.code });
1200
1455
  }
1201
1456
  async function dispatchTool(name, args, deps2) {
1202
1457
  const content = stringArg(args, "content");
1203
- if (content === null) {
1204
- return usageError("missing required 'content' argument (the message text to operate on)");
1205
- }
1458
+ if (content === null) return usageError();
1206
1459
  const runDeps = inlineDeps(content);
1207
1460
  const format = stringArg(args, "format");
1208
1461
  const fmtFlag = format !== null ? ["--format", format] : [];
@@ -1221,13 +1474,13 @@ async function dispatchTool(name, args, deps2) {
1221
1474
  return toToolResult(await convertCommand(["-", "--to", to, "--json"], runDeps, VALUE_FREE));
1222
1475
  }
1223
1476
  default:
1224
- return usageError(`unknown tool '${name}'`);
1477
+ return usageError();
1225
1478
  }
1226
1479
  } catch (e) {
1227
1480
  return internalError(e);
1228
1481
  }
1229
1482
  }
1230
- var CONTENT_PROP, FORMAT_PROP, TOOL_DEFS;
1483
+ var CONTENT_PROP, FORMAT_PROP, OUTCOME_PROPS, OUTCOME_REQUIRED, TOOL_DEFS;
1231
1484
  var init_tools = __esm({
1232
1485
  "src/mcp/tools.ts"() {
1233
1486
  init_convert();
@@ -1250,39 +1503,108 @@ var init_tools = __esm({
1250
1503
  description: "Optional format override; omit to autodetect by content. A format whose parser does not support the requested operation returns a value-free CLI_FORMAT_UNSUPPORTED (never a fake)."
1251
1504
  }
1252
1505
  };
1506
+ OUTCOME_PROPS = {
1507
+ ok: {
1508
+ type: "boolean",
1509
+ description: "True iff the tool call produced data. A negative verdict about the message is still true (the tool worked); only a call that produced no data at all is false."
1510
+ },
1511
+ status: {
1512
+ type: "string",
1513
+ enum: ["success", "verdict", "failed"],
1514
+ description: "The outcome to branch on without parsing text: 'success' the operation completed cleanly, 'verdict' the tool ran and reports a negative finding about the message, 'failed' the call produced no data (a usage mistake, unparseable input, an unavailable parser, an internal error)."
1515
+ },
1516
+ exit: {
1517
+ type: "integer",
1518
+ enum: Object.values(EXIT),
1519
+ description: "The exit code the underlying command resolved to, from the documented exit-code contract."
1520
+ },
1521
+ code: {
1522
+ type: "string",
1523
+ enum: Object.values(CLI_CODES),
1524
+ description: "The stable diagnostic code on a failed call, drawn from a fixed registry so it never carries an input value. Absent on a call that produced data."
1525
+ }
1526
+ };
1527
+ OUTCOME_REQUIRED = ["ok", "status", "exit"];
1253
1528
  TOOL_DEFS = [
1254
1529
  {
1255
1530
  name: "parse",
1531
+ title: "Parse a healthcare message",
1256
1532
  description: "Parse a healthcare message (HL7 v2 or FHIR R4) to typed JSON. Format is autodetected by content. Returns the parsed model plus value-free warnings.",
1257
1533
  inputSchema: {
1258
1534
  type: "object",
1259
1535
  properties: { ...CONTENT_PROP, ...FORMAT_PROP },
1260
1536
  required: ["content"],
1261
1537
  additionalProperties: false
1262
- }
1538
+ },
1539
+ outputSchema: outputSchema({
1540
+ type: "object",
1541
+ description: "The parse payload, present on a call that produced data. A single message parses to format + model + warnings. A multi-record input (an MLLP stream) parses to `records`, one entry per record: a parsed record, or a value-free per-record error code.",
1542
+ properties: {
1543
+ format: { type: "string", description: "The format the input was parsed as." },
1544
+ model: { description: "The wrapped library's parsed model, verbatim: the data channel." },
1545
+ warnings: {
1546
+ type: "array",
1547
+ description: "Value-free parse warnings: a stable code plus positional context."
1548
+ },
1549
+ records: {
1550
+ type: "array",
1551
+ description: "One entry per record of a multi-record input, in stream order."
1552
+ }
1553
+ },
1554
+ additionalProperties: false
1555
+ })
1263
1556
  },
1264
1557
  {
1265
1558
  name: "validate",
1559
+ title: "Validate a healthcare message",
1266
1560
  description: "Validate a message and carry the verdict: ok=true with a valid result, or a result reporting value-free findings. The parsed-but-invalid verdict is a successful call (not a tool error).",
1267
1561
  inputSchema: {
1268
1562
  type: "object",
1269
1563
  properties: { ...CONTENT_PROP, ...FORMAT_PROP },
1270
1564
  required: ["content"],
1271
1565
  additionalProperties: false
1272
- }
1566
+ },
1567
+ outputSchema: outputSchema({
1568
+ type: "object",
1569
+ description: "The validation payload, present on a call that produced data (a valid result and a negative verdict alike). Value-free throughout: findings are codes, severities and positional locators, never a field value.",
1570
+ properties: {
1571
+ format: { type: "string", description: "The format the input was validated as." },
1572
+ valid: {
1573
+ type: "boolean",
1574
+ description: "The verdict: false means parsed-but-non-conformant, not a failed call."
1575
+ },
1576
+ findings: {
1577
+ type: "array",
1578
+ description: "Value-free findings: a stable code, a severity, and a positional locator."
1579
+ }
1580
+ },
1581
+ required: ["format", "valid", "findings"],
1582
+ additionalProperties: false
1583
+ })
1273
1584
  },
1274
1585
  {
1275
1586
  name: "inspect",
1587
+ title: "Inspect a healthcare message's structure",
1276
1588
  description: "Return a value-free structural summary of a message: its type, segment/entry counts, and a warning/issue count. Never includes a field value.",
1277
1589
  inputSchema: {
1278
1590
  type: "object",
1279
1591
  properties: { ...CONTENT_PROP, ...FORMAT_PROP },
1280
1592
  required: ["content"],
1281
1593
  additionalProperties: false
1282
- }
1594
+ },
1595
+ outputSchema: outputSchema({
1596
+ type: "object",
1597
+ description: "The value-free structural summary, present on a call that produced data. `format` names the shape of the rest, which is that format's own counts and structural type codes (a message type, a resource type, a transaction-set id): classification, never a value.",
1598
+ properties: {
1599
+ format: { type: "string", description: "The format the summary describes." }
1600
+ },
1601
+ required: ["format"],
1602
+ additionalProperties: true
1603
+ })
1283
1604
  },
1284
1605
  {
1285
1606
  name: "convert",
1607
+ title: "Convert HL7 v2 to a FHIR R4 Bundle",
1286
1608
  description: "Convert an HL7 v2 message to a FHIR R4 Bundle via @cosyte/transform. Returns the converted Bundle; an error-severity conversion issue is reported with ok=false.",
1287
1609
  inputSchema: {
1288
1610
  type: "object",
@@ -1296,7 +1618,21 @@ var init_tools = __esm({
1296
1618
  },
1297
1619
  required: ["content"],
1298
1620
  additionalProperties: false
1299
- }
1621
+ },
1622
+ outputSchema: outputSchema({
1623
+ type: "object",
1624
+ description: "The conversion payload, present on a call that produced data. An error-severity finding is a negative verdict (the Bundle is still returned), not a failed call.",
1625
+ properties: {
1626
+ format: { type: "string", description: "The conversion target the Bundle is in." },
1627
+ bundle: { type: "object", description: "The converted FHIR R4 Bundle: the data channel." },
1628
+ findings: {
1629
+ type: "array",
1630
+ description: "Value-free conversion findings: a code, a severity, and a positional locator."
1631
+ }
1632
+ },
1633
+ required: ["format", "bundle", "findings"],
1634
+ additionalProperties: false
1635
+ })
1300
1636
  }
1301
1637
  ];
1302
1638
  }
@@ -1314,12 +1650,21 @@ function createMcpServer() {
1314
1650
  server.setRequestHandler(ListToolsRequestSchema, () => ({
1315
1651
  tools: TOOL_DEFS.map((t) => ({
1316
1652
  name: t.name,
1653
+ title: t.title,
1317
1654
  description: t.description,
1318
1655
  inputSchema: {
1319
1656
  type: t.inputSchema.type,
1320
1657
  properties: { ...t.inputSchema.properties },
1321
1658
  ...t.inputSchema.required !== void 0 ? { required: [...t.inputSchema.required] } : {},
1322
1659
  ...t.inputSchema.additionalProperties !== void 0 ? { additionalProperties: t.inputSchema.additionalProperties } : {}
1660
+ },
1661
+ // Every advertised tool publishes its result schema; a client validates a reply against it
1662
+ // instead of pattern-matching the text block. Declared, so it is copied unconditionally.
1663
+ outputSchema: {
1664
+ type: t.outputSchema.type,
1665
+ properties: { ...t.outputSchema.properties },
1666
+ required: [...t.outputSchema.required],
1667
+ additionalProperties: t.outputSchema.additionalProperties
1323
1668
  }
1324
1669
  }))
1325
1670
  }));
@@ -1329,7 +1674,9 @@ function createMcpServer() {
1329
1674
  return {
1330
1675
  content: result.content.map((c) => ({ type: c.type, text: c.text })),
1331
1676
  isError: result.isError,
1332
- structuredContent: { exit: result.structuredContent.exit, ok: result.structuredContent.ok }
1677
+ // The whole structured value reaches the wire. A field-by-field copy here would silently
1678
+ // truncate the payload the called tool's own schema promises.
1679
+ structuredContent: { ...result.structuredContent }
1333
1680
  };
1334
1681
  });
1335
1682
  return server;
@@ -1347,27 +1694,73 @@ var init_server = __esm({
1347
1694
  }
1348
1695
  });
1349
1696
 
1697
+ // src/bin/cosyte.ts
1698
+ init_diagnostics();
1699
+ init_exit_codes();
1700
+
1350
1701
  // src/core/io.ts
1351
1702
  init_diagnostics();
1352
1703
  init_exit_codes();
1353
- async function readFileBytes(path) {
1704
+ init_limits();
1705
+ function noInputError(path) {
1706
+ return new CliError(
1707
+ CLI_CODES.CLI_NO_INPUT,
1708
+ EXIT.NOINPUT,
1709
+ `cannot read input file: ${path} (does it exist and is it readable?)`
1710
+ );
1711
+ }
1712
+ async function readFileBytes(path, limit = MAX_INPUT_BYTES) {
1713
+ let handle;
1354
1714
  try {
1355
- return await readFile(path);
1715
+ handle = await open(path, "r");
1356
1716
  } catch {
1357
- throw new CliError(
1358
- CLI_CODES.CLI_NO_INPUT,
1359
- EXIT.NOINPUT,
1360
- `cannot read input file: ${path} (does it exist and is it readable?)`
1361
- );
1717
+ throw noInputError(path);
1718
+ }
1719
+ try {
1720
+ const info = await handle.stat();
1721
+ if (info.isFile() && info.size > limit) throw inputTooLargeError(limit);
1722
+ return await handle.readFile();
1723
+ } catch (e) {
1724
+ if (e instanceof CliError) throw e;
1725
+ throw noInputError(path);
1726
+ } finally {
1727
+ await handle.close();
1362
1728
  }
1363
1729
  }
1364
- async function readStreamBytes(stream) {
1730
+ async function readStreamBytes(stream, limit = MAX_INPUT_BYTES) {
1365
1731
  const chunks = [];
1732
+ let total = 0;
1366
1733
  for await (const chunk of stream) {
1367
- chunks.push(typeof chunk === "string" ? Buffer.from(chunk, "utf-8") : chunk);
1734
+ const buf = typeof chunk === "string" ? Buffer.from(chunk, "utf-8") : chunk;
1735
+ total += buf.length;
1736
+ if (total > limit) throw inputTooLargeError(limit);
1737
+ chunks.push(buf);
1368
1738
  }
1369
1739
  return Buffer.concat(chunks);
1370
1740
  }
1741
+ async function* streamChunks(stream) {
1742
+ for await (const chunk of stream) {
1743
+ yield typeof chunk === "string" ? Buffer.from(chunk, "utf-8") : chunk;
1744
+ }
1745
+ }
1746
+ async function* fileChunks(path) {
1747
+ let handle;
1748
+ try {
1749
+ handle = await open(path, "r");
1750
+ } catch {
1751
+ throw noInputError(path);
1752
+ }
1753
+ try {
1754
+ const stream = handle.createReadStream();
1755
+ try {
1756
+ for await (const chunk of stream) yield chunk;
1757
+ } catch {
1758
+ throw noInputError(path);
1759
+ }
1760
+ } finally {
1761
+ await handle.close();
1762
+ }
1763
+ }
1371
1764
 
1372
1765
  // src/commands/completion.ts
1373
1766
  init_diagnostics();
@@ -1628,47 +2021,248 @@ function renderNote(result) {
1628
2021
  // src/core/run.ts
1629
2022
  init_parse();
1630
2023
 
1631
- // src/commands/redact.ts
1632
- init_diagnostics();
1633
-
1634
2024
  // src/core/deid.ts
1635
- var DEID_UNAVAILABLE_REASON = "redact/deid is not yet available. It delegates to @cosyte/deid, which the CLI does not wire yet. The CLI ships no built-in redactor: a partial Safe-Harbor scrub over only the obvious PHI loci would leave PHI behind and present a false-safety impression, so nothing is emitted. This command will produce a de-identified copy once that integration lands and is vetted.";
1636
- function deidStatus() {
1637
- return { available: false, reason: DEID_UNAVAILABLE_REASON };
2025
+ init_diagnostics();
2026
+ init_exit_codes();
2027
+ init_parsers();
2028
+ init_wrap();
2029
+ var DEID_PACKAGE = "@cosyte/deid";
2030
+ var DEID_COVERAGE = {
2031
+ hl7: "covered",
2032
+ fhir: "covered",
2033
+ x12: "covered",
2034
+ ccda: "covered",
2035
+ // The delegate's NCPDP adapter is Telecom; this CLI resolves SCRIPT.
2036
+ ncpdp: "no-adapter",
2037
+ astm: "no-adapter",
2038
+ mllp: "no-adapter",
2039
+ // Covered by the delegate, but its result is Part 10 bytes and stdout here is text.
2040
+ dicom: "unserializable"
2041
+ };
2042
+ function deidCoverage(format) {
2043
+ return DEID_COVERAGE[format];
2044
+ }
2045
+ function isDeidCovered(format) {
2046
+ return DEID_COVERAGE[format] === "covered";
2047
+ }
2048
+ function deidCoveredFormats() {
2049
+ return Object.keys(DEID_COVERAGE).filter(isDeidCovered).sort();
2050
+ }
2051
+ var DEID_BLOCKED = "blocked";
2052
+ var DEID_UNAVAILABLE_REASON = `the ${DEID_PACKAGE} de-identification library is not installed; install it to use redact/deid (it is an optional dependency). The CLI ships no built-in redactor: a partial Safe-Harbor scrub over only the obvious PHI loci would leave PHI behind and present a false-safety impression, so nothing is emitted`;
2053
+ function adapterUnavailable(format) {
2054
+ return `the ${DEID_PACKAGE}/${format} de-identification adapter could not be loaded; it needs both ${DEID_PACKAGE} and the @cosyte/${format} parser to be installed`;
2055
+ }
2056
+ function parseOrFail(format, parse) {
2057
+ try {
2058
+ return parse();
2059
+ } catch (e) {
2060
+ const code = extractStableCode(e);
2061
+ throw new CliError(
2062
+ CLI_CODES.CLI_PARSE_FAILED,
2063
+ EXIT.DATAERR,
2064
+ `the ${format} parser rejected the input${code === null ? "" : ` (${code})`}`
2065
+ );
2066
+ }
2067
+ }
2068
+ function deidentifyOrFail(run2) {
2069
+ try {
2070
+ return run2();
2071
+ } catch (e) {
2072
+ const code = extractStableCode(e);
2073
+ if (code === null) throw e;
2074
+ throw new CliError(
2075
+ CLI_CODES.CLI_DEID_INCOMPLETE,
2076
+ EXIT.INVALID,
2077
+ `the ${DEID_PACKAGE} de-identifier could not complete this input (${code}); no output was emitted`
2078
+ );
2079
+ }
2080
+ }
2081
+ function decode2(bytes) {
2082
+ return new TextDecoder("utf-8", { fatal: false }).decode(bytes);
2083
+ }
2084
+ function toRecords(manifest) {
2085
+ return manifest.map((e) => ({
2086
+ category: e.category,
2087
+ transform: e.transform,
2088
+ locus: e.locus,
2089
+ count: e.count,
2090
+ disposition: e.disposition,
2091
+ code: e.code
2092
+ }));
2093
+ }
2094
+ async function loadDeidDelegate(importCore = () => import('@cosyte/deid')) {
2095
+ const core = await loadOptionalPackage(DEID_UNAVAILABLE_REASON, importCore);
2096
+ const context = core.createDeidContext({ key: randomBytes(32) });
2097
+ return {
2098
+ label: core.OUTPUT_LABEL,
2099
+ version: core.VERSION,
2100
+ redact: async (format, bytes) => {
2101
+ switch (format) {
2102
+ case "hl7": {
2103
+ const { parseHL7 } = await import('@cosyte/hl7');
2104
+ const { deidentifyHl7 } = await loadOptionalPackage(
2105
+ adapterUnavailable("hl7"),
2106
+ () => import('@cosyte/deid/hl7')
2107
+ );
2108
+ const msg = parseOrFail("hl7", () => parseHL7(Buffer.from(bytes)));
2109
+ return deidentifyOrFail(() => {
2110
+ const { document, manifest } = deidentifyHl7(msg, { context });
2111
+ return { output: document.toString(), manifest: toRecords(manifest) };
2112
+ });
2113
+ }
2114
+ case "fhir": {
2115
+ const { parseResource, serializeResource } = await loadFhir();
2116
+ const { deidentifyFhir } = await loadOptionalPackage(
2117
+ adapterUnavailable("fhir"),
2118
+ () => import('@cosyte/deid/fhir')
2119
+ );
2120
+ const resource = parseOrFail("fhir", () => parseResource(decode2(bytes)).resource);
2121
+ return deidentifyOrFail(() => {
2122
+ const { document, manifest } = deidentifyFhir(resource, { context });
2123
+ return { output: serializeResource(document), manifest: toRecords(manifest) };
2124
+ });
2125
+ }
2126
+ case "x12": {
2127
+ const { parseX12 } = await loadOptional("x12", () => import('@cosyte/x12'));
2128
+ const { deidentifyX12 } = await loadOptionalPackage(
2129
+ adapterUnavailable("x12"),
2130
+ () => import('@cosyte/deid/x12')
2131
+ );
2132
+ const interchange = parseOrFail("x12", () => parseX12(Buffer.from(bytes)));
2133
+ return deidentifyOrFail(() => {
2134
+ const { x12, manifest } = deidentifyX12(interchange, { context });
2135
+ return { output: x12, manifest: toRecords(manifest) };
2136
+ });
2137
+ }
2138
+ case "ccda": {
2139
+ const { parseCcda, serializeCcda } = await loadOptional(
2140
+ "ccda",
2141
+ () => import('@cosyte/ccda')
2142
+ );
2143
+ const { deidentifyCcda } = await loadOptionalPackage(
2144
+ adapterUnavailable("ccda"),
2145
+ () => import('@cosyte/deid/ccda')
2146
+ );
2147
+ const doc = parseOrFail("ccda", () => parseCcda(decode2(bytes)));
2148
+ return deidentifyOrFail(() => {
2149
+ const { document, manifest } = deidentifyCcda(doc, { context });
2150
+ return { output: serializeCcda(document), manifest: toRecords(manifest) };
2151
+ });
2152
+ }
2153
+ }
2154
+ }
2155
+ };
1638
2156
  }
1639
2157
 
1640
2158
  // src/commands/redact.ts
2159
+ init_diagnostics();
1641
2160
  init_exit_codes();
2161
+ init_input();
1642
2162
  var REDACT_OPTIONS = {
1643
2163
  format: { type: "string" }
1644
2164
  };
1645
- function redactCommand(args) {
2165
+ function manifestLine(entry) {
2166
+ return `cosyte: redact: ${entry.category} ${entry.transform} ${entry.locus} x${String(entry.count)} ${entry.disposition} ${entry.code}
2167
+ `;
2168
+ }
2169
+ function attribution(delegate) {
2170
+ return `cosyte: redact: delegated to ${DEID_PACKAGE} ${delegate.version}: ${delegate.label}
2171
+ `;
2172
+ }
2173
+ function manifestLines(outcome) {
2174
+ return outcome.manifest.map(manifestLine).join("");
2175
+ }
2176
+ function countOf(outcome, disposition) {
2177
+ return outcome.manifest.filter((e) => e.disposition === disposition).length;
2178
+ }
2179
+ function tally(format, outcome) {
2180
+ const transformed = countOf(outcome, "transformed");
2181
+ const removed = countOf(outcome, "removed");
2182
+ const blocked = countOf(outcome, DEID_BLOCKED);
2183
+ return `cosyte: redact: ${format}: ${String(outcome.manifest.length)} loci acted on (${String(transformed)} transformed, ${String(removed)} removed, ${String(blocked)} blocked)
2184
+ `;
2185
+ }
2186
+ var EPHEMERAL_KEY_NOTE = "cosyte: redact: identifier surrogates are keyed with a per-invocation ephemeral key: consistent within this output, not stable across runs\n";
2187
+ async function redactCommand(args, deps2, loadDelegate = loadDeidDelegate) {
2188
+ let values;
2189
+ let positionals;
1646
2190
  try {
1647
- parseArgs({ args, options: REDACT_OPTIONS, allowPositionals: true });
2191
+ const parsed = parseArgs({ args, options: REDACT_OPTIONS, allowPositionals: true });
2192
+ values = parsed.values;
2193
+ positionals = parsed.positionals;
1648
2194
  } catch {
2195
+ return errorResult(
2196
+ new CliError(
2197
+ CLI_CODES.CLI_USAGE,
2198
+ EXIT.USAGE,
2199
+ "invalid arguments to `redact` (see `cosyte --help`)"
2200
+ )
2201
+ );
2202
+ }
2203
+ let delegate;
2204
+ try {
2205
+ delegate = await loadDelegate();
2206
+ } catch (e) {
2207
+ if (e instanceof CliError) return errorResult(e);
2208
+ throw e;
2209
+ }
2210
+ const resolved = await resolveInput(positionals[0], values.format, deps2, null);
2211
+ if (!resolved.ok) return resolved.result;
2212
+ const { format, bytes } = resolved.input;
2213
+ if (!isDeidCovered(format)) return errorResult(uncovered(format));
2214
+ let outcome;
2215
+ try {
2216
+ outcome = await delegate.redact(format, bytes);
2217
+ } catch (e) {
2218
+ if (e instanceof CliError) return errorResult(e);
2219
+ throw e;
2220
+ }
2221
+ const blocked = outcome.manifest.filter((e) => e.disposition === DEID_BLOCKED);
2222
+ if (blocked.length > 0) {
2223
+ const incomplete = new CliError(
2224
+ CLI_CODES.CLI_DEID_INCOMPLETE,
2225
+ EXIT.INVALID,
2226
+ `${DEID_PACKAGE} blocked ${String(blocked.length)} of ${String(outcome.manifest.length)} loci in this ${format} input; no output was emitted`
2227
+ );
1649
2228
  return {
1650
2229
  stdout: "",
1651
- stderr: `cosyte: ${CLI_CODES.CLI_USAGE}: invalid arguments to \`redact\` (see \`cosyte --help\`)
2230
+ stderr: attribution(delegate) + manifestLines(outcome) + tally(format, outcome) + `cosyte: ${incomplete.code}: ${incomplete.message}
1652
2231
  `,
1653
- exit: EXIT.USAGE
2232
+ exit: incomplete.exit
1654
2233
  };
1655
2234
  }
1656
- const status = deidStatus();
1657
2235
  return {
1658
- stdout: "",
1659
- stderr: `cosyte: ${CLI_CODES.CLI_NOT_IMPLEMENTED}: ${status.reason}
2236
+ stdout: `${outcome.output}
1660
2237
  `,
1661
- exit: EXIT.UNAVAILABLE
2238
+ stderr: attribution(delegate) + manifestLines(outcome) + tally(format, outcome) + EPHEMERAL_KEY_NOTE,
2239
+ exit: EXIT.OK
1662
2240
  };
1663
2241
  }
2242
+ function uncovered(format) {
2243
+ if (deidCoverage(format) === "unserializable") {
2244
+ return new CliError(
2245
+ CLI_CODES.CLI_FORMAT_UNSUPPORTED,
2246
+ EXIT.DATAERR,
2247
+ `format '${format}' cannot be redacted onto this CLI's text output channel (its de-identified form is a binary stream, so \`parse\` and \`fmt\` are unwired too)`
2248
+ );
2249
+ }
2250
+ return new CliError(
2251
+ CLI_CODES.CLI_NOT_IMPLEMENTED,
2252
+ EXIT.UNAVAILABLE,
2253
+ `${DEID_PACKAGE} does not de-identify the ${format} standard this CLI resolves, and the CLI ships no redactor of its own, so nothing is emitted`
2254
+ );
2255
+ }
1664
2256
 
1665
2257
  // src/core/run.ts
1666
2258
  init_validate();
1667
2259
  init_diagnostics();
1668
2260
  init_exit_codes();
2261
+ init_limits();
1669
2262
  init_phi();
1670
2263
  init_version();
1671
- var HELP = `cosyte: a PHI-safe developer CLI over the @cosyte/* healthcare parsers
2264
+ function helpText(limitBytes = MAX_INPUT_BYTES) {
2265
+ return `cosyte: a PHI-safe developer CLI over the @cosyte/* healthcare parsers
1672
2266
 
1673
2267
  Usage:
1674
2268
  cosyte <command> [options]
@@ -1680,7 +2274,7 @@ Commands:
1680
2274
  fmt <file|-> Canonically re-serialize via the parser's spec-clean serializer
1681
2275
  convert <file|-> Convert HL7 v2 \u2192 FHIR R4 via @cosyte/transform (use --to fhir)
1682
2276
  map-codes <cmap|-> Translate a code through a BYO FHIR ConceptMap via @cosyte/terminology
1683
- redact <file|-> De-identify a message (alias: deid): gated on @cosyte/deid, not yet wired
2277
+ redact <file|-> De-identify a message via @cosyte/deid (alias: deid): ${deidCoveredFormats().join(" ")}
1684
2278
  mcp Start the stdio MCP server (agent front door; also the cosyte-mcp bin)
1685
2279
  completion <shell> Print a shell completion script (bash | zsh | fish)
1686
2280
 
@@ -1701,22 +2295,44 @@ Common options (parse / validate / inspect / fmt):
1701
2295
  convert options:
1702
2296
  --to fhir The conversion target (required; only HL7 v2 \u2192 FHIR R4 today)
1703
2297
 
2298
+ redact / deid (de-identification is delegated to @cosyte/deid; the CLI adds no policy of its own):
2299
+ Covered formats: ${deidCoveredFormats().join(", ")}. Every other format is refused rather than
2300
+ approximated: astm, mllp and ncpdp have no adapter (69), and dicom is covered by the library but
2301
+ its de-identified form is a binary stream this text stdout cannot carry (65). If the library
2302
+ reports it could not handle an element, the run exits 1 and emits no output at all: a partial
2303
+ scrub presented as de-identified is the one outcome this command exists to prevent.
2304
+ stdout carries the de-identified document; stderr carries the library's own value-free manifest
2305
+ (category, transform, locus, count, disposition code) plus its published label and version.
2306
+ Identifier surrogates are keyed with a per-invocation ephemeral key: consistent within one
2307
+ output, not stable across runs.
2308
+
1704
2309
  map-codes options (the positional is a BYO FHIR ConceptMap; a code is not PHI):
1705
2310
  --code <code> The source code to translate (required)
1706
2311
  --system <uri> The source code system (optional; selects the ConceptMap group)
1707
2312
  --version <v> The source code system version (optional)
1708
2313
  --display <d> The source display (optional)
1709
2314
 
2315
+ Input size:
2316
+ parse reads at most ${describeByteLimit(limitBytes)} of input per invocation. A larger input is
2317
+ refused with CLI_INPUT_TOO_LARGE and the data-error exit code (65), never the internal-error
2318
+ code; split the input and re-run. Multi-record output (--ndjson, MLLP) is written record by
2319
+ record as it is parsed, so a refusal can arrive after some records have already been emitted:
2320
+ the exit code, not the output, is the signal that the run did not complete.
2321
+
1710
2322
  Exit codes:
1711
- 0 success / valid 65 data error (unparseable / undetected format)
2323
+ 0 success / valid 65 data error (unparseable / undetected format / input too large)
1712
2324
  1 invalid (validate) 66 no input (missing/unreadable file)
1713
- 2 usage error 69 unavailable (a capability is not yet built, e.g. redact)
2325
+ or de-identification incomplete (redact)
2326
+ 2 usage error 69 unavailable (a capability is not wired here, e.g. redact on a
2327
+ format @cosyte/deid has no adapter for)
1714
2328
  70 internal error
1715
2329
 
1716
2330
  PHI posture: the parsed model goes to stdout (the data channel you chose); every diagnostic on
1717
2331
  stderr is value-free (codes and positions only, never a field value) unless you pass the loud,
1718
2332
  opt-in --unsafe-show-values (which permits a bounded input excerpt in a failure diagnostic).
1719
2333
  `;
2334
+ }
2335
+ var HELP = helpText();
1720
2336
  function wantsHelp(argv) {
1721
2337
  return argv.includes("--help") || argv.includes("-h");
1722
2338
  }
@@ -1746,7 +2362,7 @@ async function run(argv, deps2) {
1746
2362
  return await mapCodesCommand(rest, deps2);
1747
2363
  case "redact":
1748
2364
  case "deid":
1749
- return redactCommand(rest);
2365
+ return await redactCommand(rest, deps2);
1750
2366
  case "completion":
1751
2367
  return completionCommand(rest);
1752
2368
  default:
@@ -1768,22 +2384,90 @@ function resolve(e) {
1768
2384
  }
1769
2385
 
1770
2386
  // src/bin/cosyte.ts
2387
+ var serverMode = process.argv[2] === "mcp";
2388
+ var stdoutOpen = true;
2389
+ var stderrOpen = true;
2390
+ var reported = false;
2391
+ process.stderr.on("error", () => {
2392
+ stderrOpen = false;
2393
+ });
2394
+ function note(line) {
2395
+ if (!stderrOpen) return;
2396
+ try {
2397
+ process.stderr.write(line);
2398
+ } catch {
2399
+ stderrOpen = false;
2400
+ }
2401
+ }
2402
+ function reportOutputClosed() {
2403
+ if (reported) return;
2404
+ reported = true;
2405
+ note(
2406
+ `cosyte: ${CLI_CODES.CLI_OUTPUT_WRITE_FAILED}: could not write to the output stream; it closed before the output was delivered
2407
+ `
2408
+ );
2409
+ process.exitCode = EXIT.IOERR;
2410
+ }
2411
+ process.stdout.on("error", () => {
2412
+ stdoutOpen = false;
2413
+ if (!serverMode) return;
2414
+ reportOutputClosed();
2415
+ process.stdin.destroy();
2416
+ });
2417
+ var produced = false;
2418
+ var delivered = true;
2419
+ var queued = 0;
2420
+ var onFlushed;
2421
+ function emit(chunk) {
2422
+ produced = true;
2423
+ if (!stdoutOpen) {
2424
+ delivered = false;
2425
+ return;
2426
+ }
2427
+ queued += 1;
2428
+ process.stdout.write(chunk, (err) => {
2429
+ if (err !== null && err !== void 0) delivered = false;
2430
+ queued -= 1;
2431
+ if (queued > 0 || onFlushed === void 0) return;
2432
+ const resume = onFlushed;
2433
+ onFlushed = void 0;
2434
+ resume();
2435
+ });
2436
+ }
2437
+ function flushed() {
2438
+ if (queued === 0) return Promise.resolve();
2439
+ return new Promise((resolve2) => {
2440
+ onFlushed = resolve2;
2441
+ });
2442
+ }
1771
2443
  var deps = {
1772
2444
  readFile: (path) => readFileBytes(path),
1773
- readStdin: () => readStreamBytes(process.stdin)
2445
+ readStdin: () => readStreamBytes(process.stdin),
2446
+ openFile: (path) => fileChunks(path),
2447
+ openStdin: () => streamChunks(process.stdin),
2448
+ writeStdout: (chunk) => {
2449
+ if (!stdoutOpen) throw new Error("stdout closed");
2450
+ emit(chunk);
2451
+ }
1774
2452
  };
1775
- if (process.argv[2] === "mcp") {
2453
+ if (serverMode) {
1776
2454
  Promise.resolve().then(() => (init_server(), server_exports)).then(({ startStdioServer: startStdioServer2 }) => startStdioServer2()).catch(() => {
1777
- process.stderr.write("cosyte: CLI_INTERNAL: the MCP server failed to start\n");
2455
+ note("cosyte: CLI_INTERNAL: the MCP server failed to start\n");
1778
2456
  process.exitCode = 70;
1779
2457
  });
1780
2458
  } else {
1781
- run(process.argv.slice(2), deps).then((result) => {
1782
- if (result.stdout) process.stdout.write(result.stdout);
1783
- if (result.stderr) process.stderr.write(result.stderr);
2459
+ run(process.argv.slice(2), deps).then(async (result) => {
2460
+ if (result.stdout !== "") emit(result.stdout);
2461
+ if (produced) process.exitCode = EXIT.IOERR;
2462
+ await flushed();
2463
+ if (produced && !(delivered && stdoutOpen)) {
2464
+ reportOutputClosed();
2465
+ return;
2466
+ }
2467
+ if (result.stderr) note(result.stderr);
1784
2468
  process.exitCode = result.exit;
1785
2469
  }).catch(() => {
1786
- process.stderr.write("cosyte: CLI_INTERNAL: an unexpected internal error occurred\n");
2470
+ note("cosyte: CLI_INTERNAL: an unexpected internal error occurred\n");
1787
2471
  process.exitCode = 70;
1788
2472
  });
1789
2473
  }