@cavelang/parser 0.1.0

package/dist/src/line.js

@@ -0,0 +1,293 @@
+/**
+ * Token-stream line parser (spec §3.2, §16).
+ *
+ * Parses the token list of one line into claim/continuation/qualifier
+ * bodies. The payload/metadata boundary is decided lexically: metadata items
+ * are unambiguous at their first token (spec §4.3 reserved characters), so
+ * everything before the first metadata token is payload.
+ */
+import { Claim, Confidence, Tag, Value, Verb } from '@cavelang/core';
+const ok = (value, problems = []) => ({ ok: true, value, problems });
+const fail = (message) => ({ ok: false, message });
+const sigmaRe = /^\((\d+(?:\.\d+)?)σ\)$/u;
+/** @returns `true` if `token` starts a metadata item (spec §4.3). */
+export const isMetaStart = (token) => token.kind === 'word' && (token.text.startsWith('@') ||
+    token.text.startsWith('#') ||
+    token.text.startsWith('+/-') ||
+    token.text === '!' ||
+    sigmaRe.test(token.text));
+/** @returns term from a single token (spec §16: atom, literal or code literal). */
+const term = (token) => {
+    switch (token.kind) {
+        case 'text':
+            return Claim.text(token.text);
+        case 'code':
+            return Claim.code(token.text);
+        default:
+            return Claim.entity(token.text);
+    }
+};
+/** @returns value from payload value tokens (attribute value or metric). */
+const valueOf = (tokens) => {
+    if (tokens.length === 0) {
+        return undefined;
+    }
+    const [head] = tokens;
+    if (tokens.length === 1 && head.kind === 'text') {
+        return Value.ofText(head.text);
+    }
+    if (tokens.length === 1 && head.kind === 'code') {
+        return Value.ofCode(head.text);
+    }
+    if (tokens.some(token => token.kind !== 'word')) {
+        return undefined;
+    }
+    return Value.parse(tokens.map(token => token.text).join(' '));
+};
+/**
+ * Parses metadata tokens (spec §6). Never fails — unrecognized tokens are
+ * reported as problems and skipped, honoring the robust-extraction goal
+ * (spec §1.6).
+ */
+export const parseMeta = (tokens, comment) => {
+    const contexts = [];
+    const tags = [];
+    const problems = [];
+    let conf;
+    let importance = false;
+    let delta;
+    let sigmaLevel;
+    let i = 0;
+    const collectValueTokens = () => {
+        const collected = [];
+        while (i < tokens.length && !isMetaStart(tokens[i]) && tokens[i].kind === 'word') {
+            collected.push(tokens[i]);
+            i += 1;
+        }
+        return collected;
+    };
+    while (i < tokens.length) {
+        const token = tokens[i];
+        i += 1;
+        if (token.kind !== 'word') {
+            problems.push(`unexpected literal ${JSON.stringify(token.text)} in metadata position`);
+            continue;
+        }
+        const word = token.text;
+        if (word === '@') {
+            const next = tokens[i];
+            const text = next?.kind === 'word' ? next.text : undefined;
+            const parsed = text === undefined ? undefined : Confidence.parse(text);
+            if (parsed === undefined) {
+                // Consume a number-like token so `@ 90` / `@ 2026` yield one clear
+                // problem instead of a cascade (§6.3: confidence must end in %).
+                if (text !== undefined && /^\d/.test(text)) {
+                    i += 1;
+                }
+                problems.push(`expected percentage ending in % after "@ "${text === undefined ? '' : `, got ${JSON.stringify(text)}`} (spec §6.3)`);
+                continue;
+            }
+            i += 1;
+            if (Number(text.slice(0, -1)) > 100) {
+                problems.push(`confidence ${text} above 100% clamped (spec §6.3)`);
+            }
+            if (conf !== undefined) {
+                problems.push('repeated confidence — only contexts and tags may repeat (spec §3.2); last one wins');
+            }
+            conf = parsed;
+            continue;
+        }
+        if (word.startsWith('@')) {
+            contexts.push(word.slice(1));
+            continue;
+        }
+        if (word === '#') {
+            problems.push('empty tag');
+            continue;
+        }
+        if (word.startsWith('#')) {
+            tags.push(Tag.parse(word.slice(1)));
+            continue;
+        }
+        if (word.startsWith('+/-')) {
+            const glued = word.slice('+/-'.length);
+            const collected = collectValueTokens().map(token_ => token_.text);
+            const raw = [...glued === '' ? [] : [glued], ...collected].join(' ');
+            if (raw === '') {
+                problems.push('expected value after +/- (spec §7.2)');
+                continue;
+            }
+            if (delta !== undefined) {
+                problems.push('repeated +/- uncertainty — only contexts and tags may repeat (spec §3.2); last one wins');
+            }
+            delta = Value.parse(raw);
+            continue;
+        }
+        const sigma = sigmaRe.exec(word);
+        if (sigma) {
+            if (sigmaLevel !== undefined) {
+                problems.push('repeated (Nσ) level — only contexts and tags may repeat (spec §3.2); last one wins');
+            }
+            sigmaLevel = Number(sigma[1]);
+            continue;
+        }
+        if (word === '!') {
+            if (importance) {
+                problems.push('repeated ! importance marker (spec §3.2)');
+            }
+            importance = true;
+            continue;
+        }
+        problems.push(`unexpected token ${JSON.stringify(word)} after payload`);
+    }
+    return {
+        meta: {
+            contexts,
+            tags,
+            importance,
+            ...conf !== undefined ? { conf } : {},
+            ...delta !== undefined ? { delta } : {},
+            ...sigmaLevel !== undefined ? { sigmaLevel } : {},
+            ...comment !== undefined ? { comment } : {}
+        },
+        problems
+    };
+};
+/** Splits tokens at the first metadata token. */
+const splitAtMeta = (tokens) => {
+    const at = tokens.findIndex(isMetaStart);
+    return at === -1 ?
+        { payload: tokens, meta: [] } :
+        { payload: tokens.slice(0, at), meta: tokens.slice(at) };
+};
+/**
+ * Parses payload tokens (spec §16 `payload`):
+ *
+ * - `attr: value` — attribute claim; the colon is canonical (spec §3.4)
+ * - legacy colonless `HAS attr value` when the value is numeric/date-like
+ * - numeric/date value — metric claim (`latency IS 30ms`; also the shape a
+ *   comparison condition canonicalizes to, e.g. `load EXCEEDS 1000 req/s`,
+ *   which keeps `WHERE value > …` filters queryable, spec §12.2)
+ * - empty — bare existence, `EXISTS` only
+ * - otherwise — relational object (single term or multi-word phrase)
+ */
+const parsePayload = (verb, tokens) => {
+    if (tokens.length === 0) {
+        return verb === 'EXISTS' ?
+            ok(Claim.none) :
+            fail(`missing object after ${verb} (the minimum line is "entity VERB object", spec §2.2)`);
+    }
+    const [head, ...rest] = tokens;
+    if (head.kind === 'word' && head.text.endsWith(':') && head.text.length > 1) {
+        const attribute = head.text.slice(0, -1);
+        const value = valueOf(rest);
+        if (value === undefined) {
+            return fail(`missing or mixed value after attribute ${JSON.stringify(attribute)} (spec §3.4)`);
+        }
+        return ok(Claim.attribute(attribute, value));
+    }
+    if (head.kind === 'word') {
+        // Glued colon: `expiry:3600s`. The payload `:` is reserved as the
+        // attribute/value separator (spec §4.3) — split it, keep a diagnostic.
+        const colonAt = head.text.indexOf(':');
+        if (colonAt > 0 && colonAt < head.text.length - 1) {
+            const attribute = head.text.slice(0, colonAt);
+            const value = valueOf([{ kind: 'word', text: head.text.slice(colonAt + 1) }, ...rest]);
+            if (value !== undefined) {
+                return ok(Claim.attribute(attribute, value), [`glued attribute colon; canonical is "${attribute}: ${value.raw}" — quote the object for a literal : (spec §3.4, §4.3)`]);
+            }
+        }
+    }
+    if (verb === 'HAS' && rest.length > 0 && head.kind === 'word') {
+        const value = valueOf(rest);
+        if (value !== undefined && (value.kind === 'number' || value.kind === 'date')) {
+            return ok(Claim.attribute(head.text, value), [`legacy colonless attribute form; canonical is "${head.text}: ${value.raw}" (spec §3.4)`]);
+        }
+    }
+    {
+        const value = valueOf(tokens);
+        if (value !== undefined && (value.kind === 'number' || value.kind === 'date')) {
+            return ok(Claim.metric(value));
+        }
+    }
+    if (tokens.length === 1) {
+        return ok(Claim.relation(term(head)));
+    }
+    if (tokens.every(token => token.kind === 'word')) {
+        return ok(Claim.relation(Claim.entity(tokens.map(token => token.text).join(' '))));
+    }
+    return fail('object phrase mixes words and literals — quote the whole object (spec §4.3)');
+};
+/** Parses `verb [NOT] payload metadata` — a continuation body (spec §8.3). */
+export const parseBody = (tokens, comment) => {
+    const [head, ...rest] = tokens;
+    if (head === undefined || head.kind !== 'word' || !Verb.isVerbToken(head.text)) {
+        return fail(`expected an UPPERCASE verb, got ${JSON.stringify(head?.text ?? 'end of line')}`);
+    }
+    const verb = head.text;
+    const negated = rest[0]?.kind === 'word' && rest[0].text === 'NOT';
+    const afterNot = negated ? rest.slice(1) : rest;
+    const { payload, meta } = splitAtMeta(afterNot);
+    const payloadResult = parsePayload(verb, payload);
+    if (!payloadResult.ok) {
+        return payloadResult;
+    }
+    const parsedMeta = parseMeta(meta, comment);
+    return ok({ verb, negated, payload: payloadResult.value, meta: parsedMeta.meta }, [...payloadResult.problems, ...parsedMeta.problems]);
+};
+/** Parses `subject verb [NOT] payload metadata` — a full claim line (spec §16). */
+export const parseClaim = (tokens, comment) => {
+    const [head, ...rest] = tokens;
+    if (head === undefined) {
+        return fail('empty claim line');
+    }
+    if (isMetaStart(head)) {
+        return fail(`expected a subject, got ${JSON.stringify(head.text)}`);
+    }
+    const body = parseBody(rest, comment);
+    if (!body.ok) {
+        return body;
+    }
+    return ok({ subject: term(head), ...body.value }, body.problems);
+};
+const isComparisonOp = (text) => ['>', '<', '>=', '<=', '=', '!='].includes(text);
+/**
+ * Parses a qualifier payload (spec §8.2): a full claim, a comparison, or a
+ * bare entity — tried in that order.
+ */
+export const parseQualifierPayload = (tokens, comment) => {
+    const negated = tokens[0]?.kind === 'word' && tokens[0].text === 'NOT';
+    const rest = negated ? tokens.slice(1) : tokens;
+    const [head, second] = rest;
+    if (head === undefined) {
+        return fail('empty qualifier payload');
+    }
+    if (second?.kind === 'word' &&
+        Verb.isVerbToken(second.text) &&
+        second.text !== 'NOT' &&
+        !isMetaStart(head)) {
+        const claim = parseClaim(rest, comment);
+        if (!claim.ok) {
+            return claim;
+        }
+        return ok({ kind: 'claim', negated, claim: claim.value }, claim.problems);
+    }
+    if (second?.kind === 'word' && isComparisonOp(second.text)) {
+        const { payload, meta } = splitAtMeta(rest.slice(2));
+        const value = valueOf(payload);
+        if (value === undefined) {
+            return fail(`missing value after comparison operator ${second.text}`);
+        }
+        const parsedMeta = parseMeta(meta, comment);
+        return ok({ kind: 'comparison', negated, left: term(head), op: second.text, value, meta: parsedMeta.meta }, parsedMeta.problems);
+    }
+    if (isMetaStart(head)) {
+        return fail(`expected qualifier payload, got ${JSON.stringify(head.text)}`);
+    }
+    const parsedMeta = parseMeta(rest.slice(1), comment);
+    if (rest.length > 1 && !isMetaStart(rest[1])) {
+        return fail('cannot parse qualifier payload as claim, comparison or entity');
+    }
+    return ok({ kind: 'entity', negated, term: term(head), meta: parsedMeta.meta }, parsedMeta.problems);
+};
+//# sourceMappingURL=line.js.map

package/dist/src/line.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"line.js","sourceRoot":"","sources":["../../src/line.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,EAAE,KAAK,EAAE,UAAU,EAAE,GAAG,EAAE,KAAK,EAAE,IAAI,EAAE,MAAM,gBAAgB,CAAA;AAQpE,MAAM,EAAE,GAAG,CAAI,KAAQ,EAAE,WAA8B,EAAE,EAAa,EAAE,CACtE,CAAC,EAAE,EAAE,EAAE,IAAI,EAAE,KAAK,EAAE,QAAQ,EAAE,CAAC,CAAA;AAEjC,MAAM,IAAI,GAAG,CAAC,OAAe,EAAiB,EAAE,CAC9C,CAAC,EAAE,EAAE,EAAE,KAAK,EAAE,OAAO,EAAE,CAAC,CAAA;AAE1B,MAAM,OAAO,GAAG,yBAAyB,CAAA;AAEzC,qEAAqE;AACrE,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,KAAY,EAAW,EAAE,CACnD,KAAK,CAAC,IAAI,KAAK,MAAM,IAAI,CACvB,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;IAC1B,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;IAC1B,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC;IAC5B,KAAK,CAAC,IAAI,KAAK,GAAG;IAClB,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CACzB,CAAA;AAEH,mFAAmF;AACnF,MAAM,IAAI,GAAG,CAAC,KAAY,EAAc,EAAE;IACxC,QAAQ,KAAK,CAAC,IAAI,EAAE,CAAC;QACnB,KAAK,MAAM;YACT,OAAO,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;QAC/B,KAAK,MAAM;YACT,OAAO,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;QAC/B;YACE,OAAO,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;IACnC,CAAC;AACH,CAAC,CAAA;AAED,4EAA4E;AAC5E,MAAM,OAAO,GAAG,CAAC,MAAwB,EAAuB,EAAE;IAChE,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACxB,OAAO,SAAS,CAAA;IAClB,CAAC;IACD,MAAM,CAAC,IAAI,CAAC,GAAG,MAAM,CAAA;IACrB,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,IAAI,IAAK,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;QACjD,OAAO,KAAK,CAAC,MAAM,CAAC,IAAK,CAAC,IAAI,CAAC,CAAA;IACjC,CAAC;IACD,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,IAAI,IAAK,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;QACjD,OAAO,KAAK,CAAC,MAAM,CAAC,IAAK,CAAC,IAAI,CAAC,CAAA;IACjC,CAAC;IACD,IAAI,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,MAAM,CAAC,EAAE,CAAC;QAChD,OAAO,SAAS,CAAA;IAClB,CAAC;IACD,OAAO,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAA;AAC/D,CAAC,CAAA;AAED;;;;GAIG;AACH,MAAM,CAAC,MAAM,SAAS,GAAG,CAAC,MAAwB,EAAE,OAAgB,EAA0C,EAAE;IAC9G,MAAM,QAAQ,GAAa,EAAE,CAAA;IAC7B,MAAM,IAAI,GAAY,EAAE,CAAA;IACxB,MAAM,QAAQ,GAAa,EAAE,CAAA;IAC7B,IAAI,IAAwB,CAAA;IAC5B,IAAI,UAAU,GAAG,KAAK,CAAA;IACtB,IAAI,KAA0B,CAAA;IAC9B,IAAI,UAA8B,CAAA;IAClC,IAAI,CAAC,GAAG,CAAC,CAAA;IACT,MAAM,kBAAkB,GAAG,GAAY,EAAE;QACvC,MAAM,SAAS,GAAY,EAAE,CAAA;QAC7B,OAAO,CAAC,GAAG,MAAM,CAAC,MAAM,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,CAAC,CAAE,CAAC,IAAI,MAAM,CAAC,CAAC,CAAE,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YACnF,SAAS,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAE,CAAC,CAAA;YAC1B,CAAC,IAAI,CAAC,CAAA;QACR,CAAC;QACD,OAAO,SAAS,CAAA;IAClB,CAAC,CAAA;IACD,OAAO,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC;QACzB,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAE,CAAA;QACxB,CAAC,IAAI,CAAC,CAAA;QACN,IAAI,KAAK,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC1B,QAAQ,CAAC,IAAI,CAAC,sBAAsB,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,CAAC,uBAAuB,CAAC,CAAA;YACtF,SAAQ;QACV,CAAC;QACD,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAA;QACvB,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjB,MAAM,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,CAAA;YACtB,MAAM,IAAI,GAAG,IAAI,EAAE,IAAI,KAAK,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,SAAS,CAAA;YAC1D,MAAM,MAAM,GAAG,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,UAAU,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;YACtE,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;gBACzB,mEAAmE;gBACnE,iEAAiE;gBACjE,IAAI,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;oBAC3C,CAAC,IAAI,CAAC,CAAA;gBACR,CAAC;gBACD,QAAQ,CAAC,IAAI,CAAC,6CAA6C,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,SAAS,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,EAAE,cAAc,CAAC,CAAA;gBACnI,SAAQ;YACV,CAAC;YACD,CAAC,IAAI,CAAC,CAAA;YACN,IAAI,MAAM,CAAC,IAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,GAAG,EAAE,CAAC;gBACrC,QAAQ,CAAC,IAAI,CAAC,cAAc,IAAI,iCAAiC,CAAC,CAAA;YACpE,CAAC;YACD,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;gBACvB,QAAQ,CAAC,IAAI,CAAC,oFAAoF,CAAC,CAAA;YACrG,CAAC;YACD,IAAI,GAAG,MAAM,CAAA;YACb,SAAQ;QACV,CAAC;QACD,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;YACzB,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAA;YAC5B,SAAQ;QACV,CAAC;QACD,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjB,QAAQ,CAAC,IAAI,CAAC,WAAW,CAAC,CAAA;YAC1B,SAAQ;QACV,CAAC;QACD,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;YACzB,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;YACnC,SAAQ;QACV,CAAC;QACD,IAAI,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;YAC3B,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC,CAAA;YACtC,MAAM,SAAS,GAAG,kBAAkB,EAAE,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;YACjE,MAAM,GAAG,GAAG,CAAC,GAAG,KAAK,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,EAAE,GAAG,SAAS,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;YACpE,IAAI,GAAG,KAAK,EAAE,EAAE,CAAC;gBACf,QAAQ,CAAC,IAAI,CAAC,sCAAsC,CAAC,CAAA;gBACrD,SAAQ;YACV,CAAC;YACD,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;gBACxB,QAAQ,CAAC,IAAI,CAAC,yFAAyF,CAAC,CAAA;YAC1G,CAAC;YACD,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAA;YACxB,SAAQ;QACV,CAAC;QACD,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QAChC,IAAI,KAAK,EAAE,CAAC;YACV,IAAI,UAAU,KAAK,SAAS,EAAE,CAAC;gBAC7B,QAAQ,CAAC,IAAI,CAAC,oFAAoF,CAAC,CAAA;YACrG,CAAC;YACD,UAAU,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAA;YAC7B,SAAQ;QACV,CAAC;QACD,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjB,IAAI,UAAU,EAAE,CAAC;gBACf,QAAQ,CAAC,IAAI,CAAC,0CAA0C,CAAC,CAAA;YAC3D,CAAC;YACD,UAAU,GAAG,IAAI,CAAA;YACjB,SAAQ;QACV,CAAC;QACD,QAAQ,CAAC,IAAI,CAAC,oBAAoB,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAA;IACzE,CAAC;IACD,OAAO;QACL,IAAI,EAAE;YACJ,QAAQ;YACR,IAAI;YACJ,UAAU;YACV,GAAG,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE;YACrC,GAAG,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE;YACvC,GAAG,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE;YACjD,GAAG,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE;SAC5C;QACD,QAAQ;KACT,CAAA;AACH,CAAC,CAAA;AAED,iDAAiD;AACjD,MAAM,WAAW,GAAG,CAAC,MAAwB,EAAyD,EAAE;IACtG,MAAM,EAAE,GAAG,MAAM,CAAC,SAAS,CAAC,WAAW,CAAC,CAAA;IACxC,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC;QAChB,EAAE,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,EAAE,EAAE,CAAC,CAAC;QAC/B,EAAE,OAAO,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,IAAI,EAAE,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,EAAE,CAAA;AAC5D,CAAC,CAAA;AAED;;;;;;;;;;GAUG;AACH,MAAM,YAAY,GAAG,CAAC,IAAY,EAAE,MAAwB,EAAyB,EAAE;IACrF,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACxB,OAAO,IAAI,KAAK,QAAQ,CAAC,CAAC;YACxB,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC;YAChB,IAAI,CAAC,wBAAwB,IAAI,wDAAwD,CAAC,CAAA;IAC9F,CAAC;IACD,MAAM,CAAC,IAAI,EAAE,GAAG,IAAI,CAAC,GAAG,MAAM,CAAA;IAC9B,IAAI,IAAK,CAAC,IAAI,KAAK,MAAM,IAAI,IAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,IAAK,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC/E,MAAM,SAAS,GAAG,IAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAA;QACzC,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,CAAA;QAC3B,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,OAAO,IAAI,CAAC,0CAA0C,IAAI,CAAC,SAAS,CAAC,SAAS,CAAC,cAAc,CAAC,CAAA;QAChG,CAAC;QACD,OAAO,EAAE,CAAC,KAAK,CAAC,SAAS,CAAC,SAAS,EAAE,KAAK,CAAC,CAAC,CAAA;IAC9C,CAAC;IACD,IAAI,IAAK,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;QAC1B,kEAAkE;QAClE,uEAAuE;QACvE,MAAM,OAAO,GAAG,IAAK,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAA;QACvC,IAAI,OAAO,GAAG,CAAC,IAAI,OAAO,GAAG,IAAK,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACnD,MAAM,SAAS,GAAG,IAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,CAAA;YAC9C,MAAM,KAAK,GAAG,OAAO,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,IAAK,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,GAAG,CAAC,CAAC,EAAE,EAAE,GAAG,IAAI,CAAC,CAAC,CAAA;YACvF,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;gBACxB,OAAO,EAAE,CACP,KAAK,CAAC,SAAS,CAAC,SAAS,EAAE,KAAK,CAAC,EACjC,CAAC,wCAAwC,SAAS,KAAK,KAAK,CAAC,GAAG,wDAAwD,CAAC,CAC1H,CAAA;YACH,CAAC;QACH,CAAC;IACH,CAAC;IACD,IAAI,IAAI,KAAK,KAAK,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,IAAK,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;QAC/D,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,CAAA;QAC3B,IAAI,KAAK,KAAK,SAAS,IAAI,CAAC,KAAK,CAAC,IAAI,KAAK,QAAQ,IAAI,KAAK,CAAC,IAAI,KAAK,MAAM,CAAC,EAAE,CAAC;YAC9E,OAAO,EAAE,CACP,KAAK,CAAC,SAAS,CAAC,IAAK,CAAC,IAAI,EAAE,KAAK,CAAC,EAClC,CAAC,kDAAkD,IAAK,CAAC,IAAI,KAAK,KAAK,CAAC,GAAG,eAAe,CAAC,CAC5F,CAAA;QACH,CAAC;IACH,CAAC;IACD,CAAC;QACC,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,CAAC,CAAA;QAC7B,IAAI,KAAK,KAAK,SAAS,IAAI,CAAC,KAAK,CAAC,IAAI,KAAK,QAAQ,IAAI,KAAK,CAAC,IAAI,KAAK,MAAM,CAAC,EAAE,CAAC;YAC9E,OAAO,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAA;QAChC,CAAC;IACH,CAAC;IACD,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACxB,OAAO,EAAE,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAK,CAAC,CAAC,CAAC,CAAA;IACxC,CAAC;IACD,IAAI,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,MAAM,CAAC,EAAE,CAAC;QACjD,OAAO,EAAE,CAAC,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAA;IACpF,CAAC;IACD,OAAO,IAAI,CAAC,6EAA6E,CAAC,CAAA;AAC5F,CAAC,CAAA;AAED,8EAA8E;AAC9E,MAAM,CAAC,MAAM,SAAS,GAAG,CAAC,MAAwB,EAAE,OAAgB,EAAoB,EAAE;IACxF,MAAM,CAAC,IAAI,EAAE,GAAG,IAAI,CAAC,GAAG,MAAM,CAAA;IAC9B,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,IAAI,KAAK,MAAM,IAAI,CAAC,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QAC/E,OAAO,IAAI,CAAC,mCAAmC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,IAAI,IAAI,aAAa,CAAC,EAAE,CAAC,CAAA;IAC/F,CAAC;IACD,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,CAAA;IACtB,MAAM,OAAO,GAAG,IAAI,CAAC,CAAC,CAAC,EAAE,IAAI,KAAK,MAAM,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,KAAK,KAAK,CAAA;IAClE,MAAM,QAAQ,GAAG,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAA;IAC/C,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,GAAG,WAAW,CAAC,QAAQ,CAAC,CAAA;IAC/C,MAAM,aAAa,GAAG,YAAY,CAAC,IAAI,EAAE,OAAO,CAAC,CAAA;IACjD,IAAI,CAAC,aAAa,CAAC,EAAE,EAAE,CAAC;QACtB,OAAO,aAAa,CAAA;IACtB,CAAC;IACD,MAAM,UAAU,GAAG,SAAS,CAAC,IAAI,EAAE,OAAO,CAAC,CAAA;IAC3C,OAAO,EAAE,CACP,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,aAAa,CAAC,KAAK,EAAE,IAAI,EAAE,UAAU,CAAC,IAAI,EAAE,EACtE,CAAC,GAAG,aAAa,CAAC,QAAQ,EAAE,GAAG,UAAU,CAAC,QAAQ,CAAC,CACpD,CAAA;AACH,CAAC,CAAA;AAED,mFAAmF;AACnF,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,MAAwB,EAAE,OAAgB,EAAoB,EAAE;IACzF,MAAM,CAAC,IAAI,EAAE,GAAG,IAAI,CAAC,GAAG,MAAM,CAAA;IAC9B,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;QACvB,OAAO,IAAI,CAAC,kBAAkB,CAAC,CAAA;IACjC,CAAC;IACD,IAAI,WAAW,CAAC,IAAI,CAAC,EAAE,CAAC;QACtB,OAAO,IAAI,CAAC,2BAA2B,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;IACrE,CAAC;IACD,MAAM,IAAI,GAAG,SAAS,CAAC,IAAI,EAAE,OAAO,CAAC,CAAA;IACrC,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,CAAC;QACb,OAAO,IAAI,CAAA;IACb,CAAC;IACD,OAAO,EAAE,CAAC,EAAE,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,KAAK,EAAE,EAAE,IAAI,CAAC,QAAQ,CAAC,CAAA;AAClE,CAAC,CAAA;AAED,MAAM,cAAc,GAAG,CAAC,IAAY,EAA4B,EAAE,CAC/D,CAAC,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,CAAW,CAAC,QAAQ,CAAC,IAAwB,CAAC,CAAA;AAEjF;;;GAGG;AACH,MAAM,CAAC,MAAM,qBAAqB,GAAG,CAAC,MAAwB,EAAE,OAAgB,EAAgC,EAAE;IAChH,MAAM,OAAO,GAAG,MAAM,CAAC,CAAC,CAAC,EAAE,IAAI,KAAK,MAAM,IAAI,MAAM,CAAC,CAAC,CAAC,CAAC,IAAI,KAAK,KAAK,CAAA;IACtE,MAAM,IAAI,GAAG,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAA;IAC/C,MAAM,CAAC,IAAI,EAAE,MAAM,CAAC,GAAG,IAAI,CAAA;IAC3B,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;QACvB,OAAO,IAAI,CAAC,yBAAyB,CAAC,CAAA;IACxC,CAAC;IACD,IACE,MAAM,EAAE,IAAI,KAAK,MAAM;QACvB,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,IAAI,CAAC;QAC7B,MAAM,CAAC,IAAI,KAAK,KAAK;QACrB,CAAC,WAAW,CAAC,IAAI,CAAC,EAClB,CAAC;QACD,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,EAAE,OAAO,CAAC,CAAA;QACvC,IAAI,CAAC,KAAK,CAAC,EAAE,EAAE,CAAC;YACd,OAAO,KAAK,CAAA;QACd,CAAC;QACD,OAAO,EAAE,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,EAAE,KAAK,CAAC,QAAQ,CAAC,CAAA;IAC3E,CAAC;IACD,IAAI,MAAM,EAAE,IAAI,KAAK,MAAM,IAAI,cAAc,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC;QAC3D,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,GAAG,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAA;QACpD,MAAM,KAAK,GAAG,OAAO,CAAC,OAAO,CAAC,CAAA;QAC9B,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,OAAO,IAAI,CAAC,2CAA2C,MAAM,CAAC,IAAI,EAAE,CAAC,CAAA;QACvE,CAAC;QACD,MAAM,UAAU,GAAG,SAAS,CAAC,IAAI,EAAE,OAAO,CAAC,CAAA;QAC3C,OAAO,EAAE,CACP,EAAE,IAAI,EAAE,YAAY,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,MAAM,CAAC,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,UAAU,CAAC,IAAI,EAAE,EAChG,UAAU,CAAC,QAAQ,CACpB,CAAA;IACH,CAAC;IACD,IAAI,WAAW,CAAC,IAAI,CAAC,EAAE,CAAC;QACtB,OAAO,IAAI,CAAC,mCAAmC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;IAC7E,CAAC;IACD,MAAM,UAAU,GAAG,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,OAAO,CAAC,CAAA;IACpD,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC,EAAE,CAAC;QAC9C,OAAO,IAAI,CAAC,+DAA+D,CAAC,CAAA;IAC9E,CAAC;IACD,OAAO,EAAE,CACP,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,UAAU,CAAC,IAAI,EAAE,EACpE,UAAU,CAAC,QAAQ,CACpB,CAAA;AACH,CAAC,CAAA"}

package/dist/src/token.d.ts

@@ -0,0 +1,37 @@
+/**
+ * Line tokenizer (spec §4).
+ *
+ * Splits one line (indentation and comment already removed) into tokens:
+ * backticked code literals, double-quoted text literals, and whitespace
+ * separated words. Built on `@prelude/parser` combinators.
+ *
+ * Literals contain no escape sequences (the spec defines none) — a quoted
+ * token runs to the next matching delimiter. An unterminated delimiter
+ * degrades gracefully: the raw text is consumed as an ordinary word.
+ */
+export type Token = {
+    readonly kind: 'word';
+    readonly text: string;
+} | {
+    readonly kind: 'text';
+    readonly text: string;
+} | {
+    readonly kind: 'code';
+    readonly text: string;
+};
+export type t = Token;
+export declare const word: (text: string) => Token;
+export declare const text: (value: string) => Token;
+export declare const code: (value: string) => Token;
+/** @returns tokens of a single line. Total — any input tokenizes. */
+export declare const tokenize: (line: string) => Token[];
+/**
+ * Splits a line at the first `;` that sits outside quotes and backticks
+ * (spec §6.4). @returns the head and the trimmed comment text, `undefined`
+ * when there is no comment or it is empty.
+ */
+export declare const splitComment: (line: string) => {
+    head: string;
+    comment?: string;
+};
+//# sourceMappingURL=token.d.ts.map

package/dist/src/token.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"token.d.ts","sourceRoot":"","sources":["../../src/token.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAIH,MAAM,MAAM,KAAK,GACb;IAAE,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;CAAE,GAChD;IAAE,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;CAAE,GAChD;IAAE,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;CAAE,CAAA;AAEpD,MAAM,MAAM,CAAC,GAAG,KAAK,CAAA;AAErB,eAAO,MAAM,IAAI,GAAI,MAAM,MAAM,KAAG,KACV,CAAA;AAE1B,eAAO,MAAM,IAAI,GAAI,OAAO,MAAM,KAAG,KACJ,CAAA;AAEjC,eAAO,MAAM,IAAI,GAAI,OAAO,MAAM,KAAG,KACJ,CAAA;AAoBjC,qEAAqE;AACrE,eAAO,MAAM,QAAQ,GAAI,MAAM,MAAM,KAAG,KAAK,EAClB,CAAA;AAE3B;;;;GAIG;AACH,eAAO,MAAM,YAAY,GAAI,MAAM,MAAM,KAAG;IAAE,IAAI,EAAE,MAAM,CAAC;IAAC,OAAO,CAAC,EAAE,MAAM,CAAA;CAsB3E,CAAA"}

package/dist/src/token.js

@@ -0,0 +1,51 @@
+/**
+ * Line tokenizer (spec §4).
+ *
+ * Splits one line (indentation and comment already removed) into tokens:
+ * backticked code literals, double-quoted text literals, and whitespace
+ * separated words. Built on `@prelude/parser` combinators.
+ *
+ * Literals contain no escape sequences (the spec defines none) — a quoted
+ * token runs to the next matching delimiter. An unterminated delimiter
+ * degrades gracefully: the raw text is consumed as an ordinary word.
+ */
+import * as P from '@prelude/parser';
+export const word = (text) => ({ kind: 'word', text });
+export const text = (value) => ({ kind: 'text', text: value });
+export const code = (value) => ({ kind: 'code', text: value });
+const codeToken = P.map(P.seq(P.lit('`'), P.whileNotChars('`'), P.lit('`')), ([, body]) => code(body));
+const textToken = P.map(P.seq(P.lit('"'), P.whileNotChars('"'), P.lit('"')), ([, body]) => text(body));
+const wordToken = P.map(P.whileNotChars(' \t', 1), word);
+const anyToken = P.first(codeToken, textToken, wordToken);
+const lineTokens = P.map(P.seq(P.star(P.map(P.seq(P.ws0, anyToken), ([, token]) => token)), P.ws0), ([tokens]) => tokens);
+/** @returns tokens of a single line. Total — any input tokenizes. */
+export const tokenize = (line) => P.parse(lineTokens, line);
+/**
+ * Splits a line at the first `;` that sits outside quotes and backticks
+ * (spec §6.4). @returns the head and the trimmed comment text, `undefined`
+ * when there is no comment or it is empty.
+ */
+export const splitComment = (line) => {
+    let quote;
+    for (let i = 0; i < line.length; i++) {
+        const char = line[i];
+        if (quote !== undefined) {
+            if (char === quote) {
+                quote = undefined;
+            }
+            continue;
+        }
+        if (char === '"' || char === '`') {
+            quote = char;
+            continue;
+        }
+        if (char === ';') {
+            const comment = line.slice(i + 1).trim();
+            return comment === '' ?
+                { head: line.slice(0, i) } :
+                { head: line.slice(0, i), comment };
+        }
+    }
+    return { head: line };
+};
+//# sourceMappingURL=token.js.map

package/dist/src/token.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"token.js","sourceRoot":"","sources":["../../src/token.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAEH,OAAO,KAAK,CAAC,MAAM,iBAAiB,CAAA;AASpC,MAAM,CAAC,MAAM,IAAI,GAAG,CAAC,IAAY,EAAS,EAAE,CAC1C,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAA;AAE1B,MAAM,CAAC,MAAM,IAAI,GAAG,CAAC,KAAa,EAAS,EAAE,CAC3C,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,KAAK,EAAE,CAAC,CAAA;AAEjC,MAAM,CAAC,MAAM,IAAI,GAAG,CAAC,KAAa,EAAS,EAAE,CAC3C,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,KAAK,EAAE,CAAC,CAAA;AAEjC,MAAM,SAAS,GACb,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,aAAa,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,IAAI,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAA;AAEtF,MAAM,SAAS,GACb,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,aAAa,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,IAAI,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAA;AAEtF,MAAM,SAAS,GACb,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,aAAa,CAAC,KAAK,EAAE,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;AAExC,MAAM,QAAQ,GACZ,CAAC,CAAC,KAAK,CAAC,SAAS,EAAE,SAAS,EAAE,SAAS,CAAC,CAAA;AAE1C,MAAM,UAAU,GACd,CAAC,CAAC,GAAG,CACH,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,EAAE,QAAQ,CAAC,EAAE,CAAC,CAAC,EAAE,KAAK,CAAC,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,GAAG,CAAC,EACzE,CAAC,CAAC,MAAM,CAAC,EAAE,EAAE,CAAC,MAAM,CACrB,CAAA;AAEH,qEAAqE;AACrE,MAAM,CAAC,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAW,EAAE,CAChD,CAAC,CAAC,KAAK,CAAC,UAAU,EAAE,IAAI,CAAC,CAAA;AAE3B;;;;GAIG;AACH,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,IAAY,EAAsC,EAAE;IAC/E,IAAI,KAAyB,CAAA;IAC7B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACrC,MAAM,IAAI,GAAG,IAAI,CAAC,CAAC,CAAE,CAAA;QACrB,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;gBACnB,KAAK,GAAG,SAAS,CAAA;YACnB,CAAC;YACD,SAAQ;QACV,CAAC;QACD,IAAI,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjC,KAAK,GAAG,IAAI,CAAA;YACZ,SAAQ;QACV,CAAC;QACD,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjB,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAA;YACxC,OAAO,OAAO,KAAK,EAAE,CAAC,CAAC;gBACrB,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC;gBAC5B,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,OAAO,EAAE,CAAA;QACvC,CAAC;IACH,CAAC;IACD,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,CAAA;AACvB,CAAC,CAAA"}

package/package.json

@@ -0,0 +1,43 @@
+{
+  "name": "@cavelang/parser",
+  "version": "0.1.0",
+  "type": "module",
+  "description": "CAVE text to AST parser built on @prelude/parser combinators.",
+  "license": "CC0-1.0",
+  "author": "Mirek Rusin (https://github.com/mirek)",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/mirek/cave.git",
+    "directory": "packages/parser"
+  },
+  "engines": {
+    "node": ">=22.18"
+  },
+  "exports": {
+    ".": {
+      "types": "./dist/src/index.d.ts",
+      "default": "./dist/src/index.js"
+    }
+  },
+  "files": [
+    "dist",
+    "!dist/test",
+    "!dist/tsconfig.tsbuildinfo",
+    "src",
+    "!src/**/*.test.ts",
+    "README.md",
+    "License.md"
+  ],
+  "publishConfig": {
+    "access": "public"
+  },
+  "dependencies": {
+    "@prelude/parser": "^7.1.0",
+    "@cavelang/core": "0.1.0"
+  },
+  "scripts": {
+    "build": "tsc -b",
+    "test": "node --disable-warning=ExperimentalWarning --test 'test/*.test.ts'",
+    "typecheck": "tsc -b"
+  }
+}

package/src/ast.ts

@@ -0,0 +1,92 @@
+/**
+ * CAVE abstract syntax (spec §16).
+ *
+ * The parser turns text into `Line` values — pure syntax, before
+ * canonicalization. Inverse verbs are *not* resolved here (`packages/api
+ * PART-OF monorepo` parses with verb `PART-OF`); continuation lines keep
+ * their missing endpoint; qualifier payloads keep their surface shape.
+ * `@cavelang/canonical` applies the spec §13.4 pipeline on top of this AST.
+ */
+
+import type { Claim, Value, Tag, Context, Verb } from '@cavelang/core'
+
+/** Line metadata suffixes (spec §6), all optional. */
+export type Meta = {
+  readonly contexts: readonly Context.t[]
+  readonly tags: readonly Tag.t[]
+  /** `@ N%` as decimal; `undefined` means the default 100%. */
+  readonly conf?: number
+  readonly importance: boolean
+  /** `+/- delta` (spec §7.2). */
+  readonly delta?: Value.t
+  /** `(Nσ)` override (spec §7.2). */
+  readonly sigmaLevel?: number
+  /** `; comment` (spec §6.4). */
+  readonly comment?: string
+}
+
+/** Empty metadata. */
+export const emptyMeta: Meta =
+  { contexts: [], tags: [], importance: false }
+
+/** Claim body without a subject — what a continuation line carries (spec §8.3). */
+export type Body = {
+  readonly verb: string
+  readonly negated: boolean
+  readonly payload: Claim.Payload
+  readonly meta: Meta
+}
+
+/** Full claim: subject + body. */
+export type Full = Body & {
+  readonly subject: Claim.Term
+}
+
+/** Comparison operators accepted in qualifier payloads (`WHEN load > ~1000 req/s`). */
+export type ComparisonOp = '>' | '<' | '>=' | '<=' | '=' | '!='
+
+export const comparisonOps: readonly ComparisonOp[] =
+  ['>', '<', '>=', '<=', '=', '!=']
+
+/**
+ * Qualifier payload (spec §8.2). The grammar leaves `qualifier_payload`
+ * loose; three surface shapes cover the spec's examples:
+ *
+ * - full claim:  `WHEN memory-leak EXISTS @ 60%`
+ * - bare entity: `WHEN cache-miss`, `WHEN NOT cache/enabled`
+ * - comparison:  `WHEN load > ~1000 req/s`
+ *
+ * `negated` records a `NOT` immediately after the qualifier verb.
+ */
+export type QualifierPayload =
+  | { readonly kind: 'claim', readonly negated: boolean, readonly claim: Full }
+  | { readonly kind: 'entity', readonly negated: boolean, readonly term: Claim.Term, readonly meta: Meta }
+  | { readonly kind: 'comparison', readonly negated: boolean, readonly left: Claim.Term, readonly op: ComparisonOp, readonly value: Value.t, readonly meta: Meta }
+
+/**
+ * A parsed line. `depth` is the indentation width in characters; `parent` is
+ * the index (into the document's `lines`) of the nearest less-indented
+ * structural line above (spec §8). `raw` is the line exactly as written.
+ */
+export type Line =
+  | { readonly kind: 'blank', readonly line: number, readonly raw: string }
+  | { readonly kind: 'comment', readonly line: number, readonly raw: string, readonly text: string }
+  | { readonly kind: 'claim', readonly line: number, readonly raw: string, readonly depth: number, readonly parent?: number, readonly claim: Full }
+  | { readonly kind: 'continuation', readonly line: number, readonly raw: string, readonly depth: number, readonly parent?: number, readonly body: Body }
+  | { readonly kind: 'qualifier', readonly line: number, readonly raw: string, readonly depth: number, readonly parent?: number, readonly qualifier: Verb.Qualifier, readonly payload: QualifierPayload }
+  | { readonly kind: 'invalid', readonly line: number, readonly raw: string, readonly message: string }
+
+export type t = Line
+
+/** Parser problem tied to a 1-based line number. */
+export type Diagnostic = {
+  readonly line: number
+  readonly message: string
+  readonly raw: string
+}
+
+/** Parsed document: every input line classified, plus collected problems. */
+export type Document = {
+  readonly lines: readonly Line[]
+  readonly diagnostics: readonly Diagnostic[]
+}