/**
* Claim Parser — Extract structured claims from thought text
*
* Follows the AnchorParser pattern from src/references/anchor-parser.ts.
* Parses CLAIM:, PREMISE:, REFUTE: prefixes and turnstile (⊢) notation.
*
* @module src/multi-agent/claim-parser
*/
/**
* A structured claim extracted from thought text.
*/
export interface ExtractedClaim {
/** Type of claim: 'claim', 'premise', or 'refute' */
type: 'claim' | 'premise' | 'refute';
/** The claim content text */
content: string;
/** Whether this claim is negated (contains ¬) */
negated: boolean;
/** References to other thoughts ([S1], [S2], etc.) */
references: string[];
/** The raw matched text */
raw: string;
}
/**
* A formal derivation parsed from turnstile notation.
* Format: P₁, P₂ ⊢ C
*/
export interface Derivation {
/** Premise strings */
premises: string[];
/** Conclusion string */
conclusion: string;
/** The raw matched text */
raw: string;
}
/**
* Result of parsing a thought for claims.
*/
export interface ParseResult {
claims: ExtractedClaim[];
derivations: Derivation[];
}
/**
* Patterns for extracting claims.
*/
const CLAIM_PATTERN = /^(CLAIM|PREMISE|REFUTE):\s*(.+)$/gm;
const REF_PATTERN = /\[S\d+(?:-S\d+)?\]/g;
const NEGATION_PATTERN = /¬/;
/**
* Pattern for turnstile derivation: P₁, P₂ ⊢ C
* Supports subscript numbers (₀-₉) and regular numbers.
*/
const TURNSTILE_PATTERN = /(.+?)\s*⊢\s*(.+)/;
/**
* Parse all claims and derivations from thought text.
*
* @param text - The thought text to parse
* @returns Extracted claims and derivations
*/
export function parseClaims(text: string): ParseResult {
const claims: ExtractedClaim[] = [];
const derivations: Derivation[] = [];
// Extract CLAIM/PREMISE/REFUTE prefixed lines
let match: RegExpExecArray | null;
// Reset regex state for each call
CLAIM_PATTERN.lastIndex = 0;
while ((match = CLAIM_PATTERN.exec(text)) !== null) {
const typeStr = match[1].toLowerCase() as 'claim' | 'premise' | 'refute';
const content = match[2].trim();
const references = extractReferences(content);
const negated = NEGATION_PATTERN.test(content);
claims.push({
type: typeStr,
content,
negated,
references,
raw: match[0],
});
}
// Extract turnstile derivations
for (const line of text.split('\n')) {
const trimmed = line.trim();
const turnstileMatch = TURNSTILE_PATTERN.exec(trimmed);
if (turnstileMatch) {
const premisesStr = turnstileMatch[1].trim();
const conclusion = turnstileMatch[2].trim();
// Split premises by comma, handling subscript numbers
const premises = premisesStr
.split(/,\s*/)
.map(p => p.trim())
.filter(p => p.length > 0);
derivations.push({
premises,
conclusion,
raw: trimmed,
});
}
}
return { claims, derivations };
}
/**
* Extract [SN] references from text.
*/
function extractReferences(text: string): string[] {
const matches = text.match(REF_PATTERN);
return matches ?? [];
}
/**
* Normalize a claim content string for comparison.
* Strips whitespace, lowercases, and removes references.
*/
export function normalizeClaim(content: string): string {
return content
.replace(REF_PATTERN, '')
.replace(/\s+/g, ' ')
.trim()
.toLowerCase();
}
