Repository navigation
feat: add hasICD10Code and harden lookup input handling - #291
Merged
Merged
Conversation
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
lowlydba
commented
Nov 6, 2026
Comment on lines
+17
to
+20
| strategy: | ||
| fail-fast: false | ||
| matrix: | ||
| node-version: [22, 24] |
Contributor
Author
There was a problem hiding this comment.
note: engines allows Node 22, but CI only ran 24. npm run test:coverage also passes locally on 22.23.4.
lowlydba
commented
Nov 6, 2026
| - v3.1.0 - `ICD10_CM_RELEASE` export with the bundled fiscal year, effective date, and code count. Node.js 22+ declared in `engines`, `package.json` exported, expanded npm keywords. API, CDN, and versioning docs, plus contributing and security guides | ||
| - v3.0.0 - Data updated to the October 1, 2026 ICD-10-CM release (74,879 codes, from the FY2027 set) | ||
| - v2.0.0 - Data updated to the April 1, 2026 ICD-10-CM release (74,719 codes, from the FY2024 set). TypeScript 7 (native compiler) toolchain, ESLint replaced with oxlint + Prettier, dropped tsup for plain tsc (zero-bundler dual ESM/CJS), dataset shipped once and lazy-loaded (~50% smaller install, faster imports), fixed broken `require()` entry point | ||
| - v2.0.0 - Data updated to the April 1, 2026 ICD-10-CM release (74,719 codes, from the FY2026 set). TypeScript 7 (native compiler) toolchain, ESLint replaced with oxlint + Prettier, dropped tsup for plain tsc (zero-bundler dual ESM/CJS), dataset shipped once and lazy-loaded (~50% smaller install, faster imports), fixed broken `require()` entry point |
Contributor
Author
There was a problem hiding this comment.
note: the April 2026 release is the FY2026 set, not FY2024. Inferred from the 74,719 code count, so worth checking against the CDC files.
lowlydba
commented
Nov 6, 2026
Comment on lines
+17
to
44
| const codePattern = /^[A-Z][0-9A-Z]{2,6}$/; | ||
| const linePattern = /^(\S+)\s+(.+?)\s*$/; | ||
|
|
||
| // Function to parse the ICD file content | ||
| const parseICDFile = (content: string): ICD10Dictionary => { | ||
| const lines = content.split("\n"); // Split content into lines | ||
| const dictionary: ICD10Dictionary = {}; | ||
|
|
||
| for (const line of lines) { | ||
| const [code, ...descriptionParts] = line.split(/\s+/); // Split by whitespace (tab or spaces) | ||
| const description = descriptionParts.join(" "); // Rejoin the description parts | ||
| for (const [index, line] of content.split(/\r?\n/).entries()) { | ||
| if (line.trim() === "") { | ||
| continue; | ||
| } | ||
|
|
||
| const match = linePattern.exec(line); | ||
| const [, code, description] = match ?? []; | ||
|
|
||
| if (code && description) { | ||
| dictionary[code] = description.trim(); // Add to dictionary | ||
| if (!code || !description || !codePattern.test(code)) { | ||
| throw new Error( | ||
| `Line ${index + 1} is not "<code> <description>": ${line}`, | ||
| ); | ||
| } | ||
|
|
||
| if (Object.hasOwn(dictionary, code)) { | ||
| throw new Error(`Line ${index + 1} repeats code ${code}`); | ||
| } | ||
|
|
||
| dictionary[code] = description; | ||
| } | ||
|
|
||
| return dictionary; |
Contributor
Author
There was a problem hiding this comment.
note: the parser now throws on a malformed line or a repeated code instead of skipping or overwriting it. The pattern allows letters in the second character because FY2027 added QA-prefixed codes. Output for the current source is byte-identical.
lowlydba
commented
Nov 6, 2026
| "esModuleInterop": true, | ||
| "forceConsistentCasingInFileNames": true, | ||
| "strict": true, | ||
| "noUncheckedIndexedAccess": true, |
Contributor
Author
There was a problem hiding this comment.
note: makes the dataset lookup in src/index.cts type as string | undefined instead of relying on the declared return type.
Contributor
There was a problem hiding this comment.
🟡 Changes recommended
The parser accepts whitespace-only descriptions and writes blank values into the generated dataset, violating the trimmed, non-empty description invariant.
1 open issue
What changed in this PR
This PR adds hasICD10Code and hardens input handling in the ICD-10-CM lookup package without changing the bundled dataset.
Changes:
- Adds the boolean lookup helper, non-string input handling, and internal whitespace normalization.
- Exports the dataset JSON and documents category-header behavior.
- Strengthens parser validation, expands tests, and adds Node.js 22/24 CI coverage.
| File | Description |
|---|---|
tsconfig.json |
Enables unchecked indexed-access checking. |
tests/package.test.cjs |
Tests packaged helper, JSON export, and types. |
tests/index.test.cjs |
Covers input handling, helper behavior, and data integrity. |
tests/esm.test.mjs |
Checks the helper’s ESM export. |
src/index.mts |
Re-exports hasICD10Code for ESM consumers. |
src/index.cts |
Adds the helper and hardens normalization. |
scripts/parseICD.ts |
Validates records and anchors the output path to the script. |
README.md |
Documents new behavior and corrects release-history text. |
package.json |
Exports dataset JSON and bumps version to 3.2.0. |
package-lock.json |
Synchronizes the package version. |
.github/workflows/test.yml |
Runs CI on Node.js 22 and 24. |
Review effort: Balanced
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What and why
A review of the package turned up input-handling gaps and missing docs. The agent-tool use case in the README makes the first one realistic: a model can call
getICD10Descriptionwith a missing or non-string argument, and it threw aTypeError.undefinedfrom lookups and""fromnormalizeICD10CodehasICD10Code(code)returns a boolean, so callers don't need!== undefinedA00. 1resolvesdata/icd10.min.jsonis exported from the package. Importing it by path used to fail withERR_PACKAGE_PATH_NOT_EXPORTED, which left bundlers and edge runtimes with only the CDNE11aren't in the dataset, since the CMS file lists complete codes only. A grounding check would otherwise read a real category as a hallucinated codeVersion is 3.2.0. The data is unchanged, and regenerating it with
npm run parse-icdproduces an identical file. The one behavior change consumers can see is the whitespace stripping: inputs likeA 00.0that returnedundefinednow resolve.Testing
npm testnpm run lintAdded tests for non-string input, category headers, prototype property names, whitespace normalization,
hasICD10Code, the data export through the packed tarball, and a format check on every dataset key and description.