@regolithco/ocr 1.0.0

package/dist/browser.d.ts

@@ -0,0 +1,5 @@
+import Ocr from '@gutenye/ocr-browser';
+import { ShipRockInput, ShipMiningOrderCapture } from '@regolithco/common';
+export declare const initPaddle: () => Promise<Ocr>;
+export declare const parseShipRock: (image: string) => Promise<ShipRockInput>;
+export declare const parseShipMiningOrder: (image: string) => Promise<ShipMiningOrderCapture>;

package/dist/browser.js

@@ -0,0 +1,38 @@
+import Ocr from '@gutenye/ocr-browser';
+import { parseShipRock as parseShipRockText } from './parsers/shipRock.js';
+import { parseShipMiningOrder as parseShipMiningOrderText } from './parsers/shipMiningOrder.js';
+import log from 'loglevel';
+let ocrInstance = null;
+export const initPaddle = async () => {
+    if (!ocrInstance) {
+        // Note: For browser usage, you must provide paths to the ONNX models.
+        // Ensure these files are available in your public/assets directory.
+        ocrInstance = await Ocr.create({
+            models: {
+                detectionPath: '/assets/models/ch_PP-OCRv4_det_infer.onnx',
+                recognitionPath: '/assets/models/ch_PP-OCRv4_rec_infer.onnx',
+                dictionaryPath: '/assets/models/ppocr_keys_v1.txt',
+            },
+        });
+    }
+    return ocrInstance;
+};
+const recognizeText = async (image) => {
+    const ocr = await initPaddle();
+    const results = await ocr.detect(image);
+    // eslint-disable-next-line @typescript-eslint/no-explicit-any
+    const fullText = results.map((line) => line.text).join('\n');
+    log.debug('--- PADDLE OCR OUTPUT ---');
+    log.debug(fullText);
+    log.debug('-------------------------');
+    return fullText;
+};
+export const parseShipRock = async (image) => {
+    const text = await recognizeText(image);
+    return parseShipRockText(text);
+};
+export const parseShipMiningOrder = async (image) => {
+    const text = await recognizeText(image);
+    return parseShipMiningOrderText(text);
+};
+//# sourceMappingURL=browser.js.map

package/dist/browser.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"browser.js","sourceRoot":"","sources":["../src/browser.ts"],"names":[],"mappings":"AAAA,OAAO,GAAG,MAAM,sBAAsB,CAAA;AAEtC,OAAO,EAAE,aAAa,IAAI,iBAAiB,EAAE,MAAM,uBAAuB,CAAA;AAC1E,OAAO,EAAE,oBAAoB,IAAI,wBAAwB,EAAE,MAAM,8BAA8B,CAAA;AAC/F,OAAO,GAAG,MAAM,UAAU,CAAA;AAE1B,IAAI,WAAW,GAAe,IAAI,CAAA;AAElC,MAAM,CAAC,MAAM,UAAU,GAAG,KAAK,IAAI,EAAE;IACnC,IAAI,CAAC,WAAW,EAAE,CAAC;QACjB,sEAAsE;QACtE,oEAAoE;QACpE,WAAW,GAAG,MAAM,GAAG,CAAC,MAAM,CAAC;YAC7B,MAAM,EAAE;gBACN,aAAa,EAAE,2CAA2C;gBAC1D,eAAe,EAAE,2CAA2C;gBAC5D,cAAc,EAAE,kCAAkC;aACnD;SACF,CAAC,CAAA;IACJ,CAAC;IACD,OAAO,WAAW,CAAA;AACpB,CAAC,CAAA;AAED,MAAM,aAAa,GAAG,KAAK,EAAE,KAAa,EAAmB,EAAE;IAC7D,MAAM,GAAG,GAAG,MAAM,UAAU,EAAE,CAAA;IAC9B,MAAM,OAAO,GAAG,MAAM,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,CAAA;IACvC,8DAA8D;IAC9D,MAAM,QAAQ,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,IAAS,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IAEjE,GAAG,CAAC,KAAK,CAAC,2BAA2B,CAAC,CAAA;IACtC,GAAG,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAA;IACnB,GAAG,CAAC,KAAK,CAAC,2BAA2B,CAAC,CAAA;IAEtC,OAAO,QAAQ,CAAA;AACjB,CAAC,CAAA;AAED,MAAM,CAAC,MAAM,aAAa,GAAG,KAAK,EAAE,KAAa,EAA0B,EAAE;IAC3E,MAAM,IAAI,GAAG,MAAM,aAAa,CAAC,KAAK,CAAC,CAAA;IACvC,OAAO,iBAAiB,CAAC,IAAI,CAAC,CAAA;AAChC,CAAC,CAAA;AAED,MAAM,CAAC,MAAM,oBAAoB,GAAG,KAAK,EAAE,KAAa,EAAmC,EAAE;IAC3F,MAAM,IAAI,GAAG,MAAM,aAAa,CAAC,KAAK,CAAC,CAAA;IACvC,OAAO,wBAAwB,CAAC,IAAI,CAAC,CAAA;AACvC,CAAC,CAAA"}

package/dist/node.d.ts

@@ -0,0 +1,6 @@
+import Ocr from '@gutenye/ocr-node';
+import { ShipRockInput, ShipMiningOrderCapture } from '@regolithco/common/dist/gen/schema.types.js';
+export declare const initPaddle: () => Promise<Ocr | null>;
+export declare const recognizeText: (imagePath: string) => Promise<string>;
+export declare const parseShipMiningOrder: (imagePath: string) => Promise<ShipMiningOrderCapture>;
+export declare const parseShipRock: (imagePath: string) => Promise<ShipRockInput>;

package/dist/node.js

@@ -0,0 +1,29 @@
+import Ocr from '@gutenye/ocr-node';
+import { parseShipMiningOrder as parseShipMiningOrderText } from './parsers/shipMiningOrder.js';
+import { parseShipRock as parseShipRockText } from './parsers/shipRock.js';
+let ocrInstance = null;
+export const initPaddle = async () => {
+    if (!ocrInstance) {
+        ocrInstance = await Ocr.create();
+    }
+    return ocrInstance;
+};
+export const recognizeText = async (imagePath) => {
+    const ocr = await initPaddle();
+    const results = await ocr.detect(imagePath);
+    // eslint-disable-next-line @typescript-eslint/no-explicit-any
+    const fullText = results.map((line) => line.text).join('\n');
+    console.log('--- PADDLE OCR OUTPUT ---');
+    console.log(fullText);
+    console.log('-------------------------');
+    return fullText;
+};
+export const parseShipMiningOrder = async (imagePath) => {
+    const text = await recognizeText(imagePath);
+    return parseShipMiningOrderText(text);
+};
+export const parseShipRock = async (imagePath) => {
+    const text = await recognizeText(imagePath);
+    return parseShipRockText(text);
+};
+//# sourceMappingURL=node.js.map

package/dist/node.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"node.js","sourceRoot":"","sources":["../src/node.ts"],"names":[],"mappings":"AAAA,OAAO,GAAG,MAAM,mBAAmB,CAAA;AAEnC,OAAO,EAAE,oBAAoB,IAAI,wBAAwB,EAAE,MAAM,8BAA8B,CAAA;AAC/F,OAAO,EAAE,aAAa,IAAI,iBAAiB,EAAE,MAAM,uBAAuB,CAAA;AAE1E,IAAI,WAAW,GAAe,IAAI,CAAA;AAElC,MAAM,CAAC,MAAM,UAAU,GAAG,KAAK,IAAI,EAAE;IACnC,IAAI,CAAC,WAAW,EAAE,CAAC;QACjB,WAAW,GAAG,MAAM,GAAG,CAAC,MAAM,EAAE,CAAA;IAClC,CAAC;IACD,OAAO,WAAW,CAAA;AACpB,CAAC,CAAA;AAED,MAAM,CAAC,MAAM,aAAa,GAAG,KAAK,EAAE,SAAiB,EAAmB,EAAE;IACxE,MAAM,GAAG,GAAG,MAAM,UAAU,EAAE,CAAA;IAC9B,MAAM,OAAO,GAAG,MAAM,GAAI,CAAC,MAAM,CAAC,SAAS,CAAC,CAAA;IAC5C,8DAA8D;IAC9D,MAAM,QAAQ,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,IAAS,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IAEjE,OAAO,CAAC,GAAG,CAAC,2BAA2B,CAAC,CAAA;IACxC,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAA;IACrB,OAAO,CAAC,GAAG,CAAC,2BAA2B,CAAC,CAAA;IAExC,OAAO,QAAQ,CAAA;AACjB,CAAC,CAAA;AAED,MAAM,CAAC,MAAM,oBAAoB,GAAG,KAAK,EAAE,SAAiB,EAAmC,EAAE;IAC/F,MAAM,IAAI,GAAG,MAAM,aAAa,CAAC,SAAS,CAAC,CAAA;IAC3C,OAAO,wBAAwB,CAAC,IAAI,CAAC,CAAA;AACvC,CAAC,CAAA;AAED,MAAM,CAAC,MAAM,aAAa,GAAG,KAAK,EAAE,SAAiB,EAA0B,EAAE;IAC/E,MAAM,IAAI,GAAG,MAAM,aAAa,CAAC,SAAS,CAAC,CAAA;IAC3C,OAAO,iBAAiB,CAAC,IAAI,CAAC,CAAA;AAChC,CAAC,CAAA"}

package/dist/parser.d.ts

@@ -0,0 +1,2 @@
+export * from './parsers/shipRock.js';
+export * from './parsers/shipMiningOrder.js';

package/dist/parser.js

@@ -0,0 +1,3 @@
+export * from './parsers/shipRock.js';
+export * from './parsers/shipMiningOrder.js';
+//# sourceMappingURL=parser.js.map

package/dist/parser.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"parser.js","sourceRoot":"","sources":["../src/parser.ts"],"names":[],"mappings":"AACA,cAAc,uBAAuB,CAAC;AACtC,cAAc,8BAA8B,CAAC"}

package/dist/parsers/shipMiningOrder.d.ts

@@ -0,0 +1,2 @@
+import { ShipMiningOrderCapture } from '@regolithco/common';
+export declare const parseShipMiningOrder: (ocrText: string) => ShipMiningOrderCapture;

package/dist/parsers/shipMiningOrder.js

@@ -0,0 +1,258 @@
+import { RefineryEnum, RefineryMethodEnum, ShipOreEnum, } from '@regolithco/common';
+import { cleanLine, findBestMatch } from './utils.js';
+export const parseShipMiningOrder = (ocrText) => {
+    const lines = ocrText
+        .split('\n')
+        .map(cleanLine)
+        .filter((l) => l.length > 0);
+    let refinery;
+    let method;
+    let processDurationS;
+    const shipOres = [];
+    const expenses = [];
+    const allRefineries = Object.values(RefineryEnum);
+    const allMethods = Object.values(RefineryMethodEnum);
+    const allOres = Object.values(ShipOreEnum);
+    // Pre-calculate partial names for fuzzy matching
+    // We derive these directly from the Enum keys to avoid external dependencies
+    const refineryReverseLookup = {};
+    for (const ref of allRefineries) {
+        const enumVal = ref.toString();
+        // Strategy 1: The full enum value (normalized elsewhere, but good to have)
+        refineryReverseLookup[enumVal.toUpperCase()] = ref;
+        // Strategy 2: Split by underscore (e.g. PYRO_CHECKMATE -> PYRO, CHECKMATE)
+        const parts = enumVal.split('_');
+        for (const part of parts) {
+            if (part.length > 3) {
+                // e.g. CHECKMATE
+                refineryReverseLookup[part.toUpperCase()] = ref;
+            }
+            // Special case for L-points like CRU_L1 -> L1 (too short) but CRU_L1 -> CRU
+            // Actually L1 is not unique. combining CRU and L1 is CRU-L1.
+        }
+        // Strategy 3: Handle typical "System_Station" pattern
+        // If it has underscores, the last part is often the unique station name
+        if (parts.length > 1) {
+            // e.g. PYRO_CHECKMATE -> CHECKMATE
+            const lastPart = parts[parts.length - 1];
+            if (lastPart.length > 3) {
+                refineryReverseLookup[lastPart.toUpperCase()] = ref;
+            }
+        }
+        // Strategy 4: Cleaned "Human" version approx (CRU_L1 -> CRUL1)
+        const normalized = enumVal.replace(/_/g, '').replace(/I/g, '1').replace(/L/g, '1');
+        if (normalized.length > 3) {
+            refineryReverseLookup[normalized.toUpperCase()] = ref;
+        }
+    }
+    // State for parsing ores
+    let inMaterialsSection = false;
+    for (let idx = 0; idx < lines.length; idx++) {
+        const line = lines[idx];
+        const lineUpper = line.toUpperCase();
+        // 1. Refinery
+        if (!refinery) {
+            // Try simple inclusion of any key in our reverse lookup
+            for (const [key, ref] of Object.entries(refineryReverseLookup)) {
+                // Check if LINE includes KEY (standard) or KEY includes LINE (if line is short/fragmented)
+                if (lineUpper.includes(key) || (key.length > 4 && key.includes(lineUpper))) {
+                    refinery = ref;
+                    break;
+                }
+            }
+            // Fallback to strict Enum matching if above failed
+            if (!refinery) {
+                const normalizedLine = lineUpper.replace(/\s/g, '').replace(/-/g, '').replace(/I/g, '1').replace(/L/g, '1');
+                if (normalizedLine.length > 3) {
+                    for (const ref of allRefineries) {
+                        const normalizedRef = ref.toString().replace(/_/g, '').replace(/I/g, '1').replace(/L/g, '1');
+                        if (normalizedLine.includes(normalizedRef) || normalizedRef.includes(normalizedLine)) {
+                            refinery = ref;
+                            break;
+                        }
+                    }
+                }
+            }
+        }
+        // 2. Method
+        if (!method) {
+            const bestMethod = findBestMatch(lineUpper, allMethods, 3);
+            if (bestMethod) {
+                method = bestMethod;
+            }
+            else {
+                for (const m of allMethods) {
+                    const humanM = m.replace(/_/g, ' ');
+                    if (lineUpper.includes(humanM) || lineUpper.includes(m)) {
+                        method = m;
+                        break;
+                    }
+                    if (lineUpper.length >= humanM.length) {
+                        for (let i = 0; i <= lineUpper.length - humanM.length; i++) {
+                            const substr = lineUpper.substring(i, i + humanM.length);
+                            const dist = findBestMatch(substr, [m], 3);
+                            if (dist) {
+                                method = m;
+                                break;
+                            }
+                        }
+                    }
+                    if (method)
+                        break;
+                }
+            }
+        }
+        // 3. Process Time
+        if (!processDurationS && (lineUpper.includes('PROCESS TIME') || lineUpper.includes('TIME'))) {
+            let d = 0;
+            let h = 0;
+            let m = 0;
+            let found = false;
+            const dMatch = lineUpper.match(/(\d+)\s*D/);
+            if (dMatch) {
+                d = parseInt(dMatch[1], 10);
+                found = true;
+            }
+            const hMatch = lineUpper.match(/(\d+)\s*H/);
+            if (hMatch) {
+                h = parseInt(hMatch[1], 10);
+                found = true;
+            }
+            const mMatch = lineUpper.match(/(\d+)\s*M/);
+            if (mMatch) {
+                m = parseInt(mMatch[1], 10);
+                found = true;
+            }
+            if (found) {
+                processDurationS = d * 86400 + h * 3600 + m * 60;
+            }
+        }
+        // 3.5. Expenses
+        const feeMappings = {
+            'TOTAL COST': 'Refinery Fee',
+            'REFINERY FEE': 'Refinery Fee',
+            'SERVICE FEE': 'Service Fee',
+        };
+        const fees = Object.keys(feeMappings);
+        for (const key of fees) {
+            if (lineUpper.includes(key)) {
+                const matches = lineUpper.match(/(\d[\d,.]*)/g);
+                if (matches) {
+                    const rawAmount = matches[matches.length - 1].replace(/,/g, '');
+                    const amount = parseInt(rawAmount);
+                    if (!isNaN(amount)) {
+                        const mappedName = feeMappings[key];
+                        if (!expenses.find((e) => e.name === mappedName)) {
+                            expenses.push({
+                                __typename: 'WorkOrderExpense',
+                                name: mappedName,
+                                amount,
+                                ownerScName: 'Unknown',
+                            });
+                        }
+                    }
+                }
+            }
+        }
+        // 4. Ores Section Management
+        let exitingSection = false;
+        if (!inMaterialsSection) {
+            if (lineUpper.includes('MATERIALS SELECTED') ||
+                (lineUpper.includes('YIELD') && lineUpper.includes('REFINE') && !lineUpper.includes('TO REFINE')) ||
+                (lineUpper.includes('OTY') && lineUpper.includes('YIELD'))) {
+                inMaterialsSection = true;
+                continue;
+            }
+        }
+        else {
+            // Check for exit conditions but allow processing of current line
+            if (lineUpper.includes('TOTAL COST') ||
+                lineUpper.includes('USER DETAILS') ||
+                lineUpper.includes('PROCESSING TIME') ||
+                lineUpper.includes('Select Storage Option'.toUpperCase())) {
+                exitingSection = true;
+            }
+        }
+        if (inMaterialsSection) {
+            const words = lineUpper.split(/\s+/);
+            let oreInLine = null;
+            let oreIndex = -1;
+            for (let i = 0; i < words.length; i++) {
+                const cleanW = words[i].split('(')[0].replace(/[^A-Z]/g, '');
+                if (cleanW.length < 3)
+                    continue;
+                const match = findBestMatch(cleanW, allOres, 1);
+                if (match) {
+                    oreInLine = match;
+                    oreIndex = i;
+                    break;
+                }
+            }
+            if (oreInLine) {
+                let numbers = [];
+                for (let i = oreIndex + 1; i < words.length; i++) {
+                    if (words[i].includes('%'))
+                        continue;
+                    const raw = words[i].replace(/[^0-9.]/g, '');
+                    if (raw.length > 0) {
+                        const val = parseFloat(raw);
+                        if (!isNaN(val))
+                            numbers.push(val);
+                    }
+                }
+                // Orphan check: If not enough numbers, check previous line
+                if (numbers.length < 2 && idx > 0) {
+                    const prevLine = lines[idx - 1].toUpperCase();
+                    // Ensure previous line was not already used (simple check: does it have an Ore?)
+                    // We re-parse previous line briefly to see if it had an ID'd ore
+                    let prevLineHasOre = false;
+                    const prevWords = prevLine.split(/\s+/);
+                    for (const w of prevWords) {
+                        const c = w.split('(')[0].replace(/[^A-Z]/g, '');
+                        if (c.length >= 3 && findBestMatch(c, allOres, 1)) {
+                            prevLineHasOre = true;
+                            break;
+                        }
+                    }
+                    if (!prevLineHasOre) {
+                        // Try to extract numbers from previous line
+                        const prevNumbers = [];
+                        for (const w of prevWords) {
+                            if (w.includes('%'))
+                                continue;
+                            const raw = w.replace(/[^0-9.]/g, '');
+                            if (raw.length > 0) {
+                                const val = parseFloat(raw);
+                                if (!isNaN(val))
+                                    prevNumbers.push(val);
+                            }
+                        }
+                        if (prevNumbers.length >= 2) {
+                            numbers = prevNumbers;
+                        }
+                    }
+                }
+                if (numbers.length >= 2) {
+                    const yieldVal = numbers[1];
+                    shipOres.push({
+                        __typename: 'RefineryRowCapture',
+                        ore: oreInLine,
+                        yield: Math.floor(yieldVal),
+                    });
+                }
+            }
+            if (exitingSection) {
+                inMaterialsSection = false;
+            }
+        }
+    }
+    return {
+        __typename: 'ShipMiningOrderCapture',
+        refinery,
+        method,
+        processDurationS,
+        expenses: expenses.length > 0 ? expenses : undefined,
+        shipOres,
+    };
+};
+//# sourceMappingURL=shipMiningOrder.js.map

package/dist/parsers/shipMiningOrder.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"shipMiningOrder.js","sourceRoot":"","sources":["../../src/parsers/shipMiningOrder.ts"],"names":[],"mappings":"AAAA,OAAO,EACL,YAAY,EACZ,kBAAkB,EAClB,WAAW,GAIZ,MAAM,oBAAoB,CAAA;AAC3B,OAAO,EAAE,SAAS,EAAE,aAAa,EAAE,MAAM,YAAY,CAAA;AAErD,MAAM,CAAC,MAAM,oBAAoB,GAAG,CAAC,OAAe,EAA0B,EAAE;IAC9E,MAAM,KAAK,GAAG,OAAO;SAClB,KAAK,CAAC,IAAI,CAAC;SACX,GAAG,CAAC,SAAS,CAAC;SACd,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAA;IAE9B,IAAI,QAAkC,CAAA;IACtC,IAAI,MAAsC,CAAA;IAC1C,IAAI,gBAAoC,CAAA;IACxC,MAAM,QAAQ,GAAyB,EAAE,CAAA;IACzC,MAAM,QAAQ,GAAuB,EAAE,CAAA;IAEvC,MAAM,aAAa,GAAG,MAAM,CAAC,MAAM,CAAC,YAAY,CAAC,CAAA;IACjD,MAAM,UAAU,GAAG,MAAM,CAAC,MAAM,CAAC,kBAAkB,CAAC,CAAA;IACpD,MAAM,OAAO,GAAG,MAAM,CAAC,MAAM,CAAC,WAAW,CAAC,CAAA;IAE1C,iDAAiD;IACjD,6EAA6E;IAC7E,MAAM,qBAAqB,GAAiC,EAAE,CAAA;IAE9D,KAAK,MAAM,GAAG,IAAI,aAAa,EAAE,CAAC;QAChC,MAAM,OAAO,GAAG,GAAG,CAAC,QAAQ,EAAE,CAAA;QAE9B,2EAA2E;QAC3E,qBAAqB,CAAC,OAAO,CAAC,WAAW,EAAE,CAAC,GAAG,GAAG,CAAA;QAElD,2EAA2E;QAC3E,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,GAAG,CAAC,CAAA;QAChC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACpB,iBAAiB;gBACjB,qBAAqB,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,GAAG,GAAG,CAAA;YACjD,CAAC;YACD,4EAA4E;YAC5E,6DAA6D;QAC/D,CAAC;QAED,sDAAsD;QACtD,wEAAwE;QACxE,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACrB,mCAAmC;YACnC,MAAM,QAAQ,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAA;YACxC,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACxB,qBAAqB,CAAC,QAAQ,CAAC,WAAW,EAAE,CAAC,GAAG,GAAG,CAAA;YACrD,CAAC;QACH,CAAC;QAED,+DAA+D;QAC/D,MAAM,UAAU,GAAG,OAAO,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,GAAG,CAAC,CAAA;QAClF,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC1B,qBAAqB,CAAC,UAAU,CAAC,WAAW,EAAE,CAAC,GAAG,GAAG,CAAA;QACvD,CAAC;IACH,CAAC;IAED,yBAAyB;IACzB,IAAI,kBAAkB,GAAG,KAAK,CAAA;IAE9B,KAAK,IAAI,GAAG,GAAG,CAAC,EAAE,GAAG,GAAG,KAAK,CAAC,MAAM,EAAE,GAAG,EAAE,EAAE,CAAC;QAC5C,MAAM,IAAI,GAAG,KAAK,CAAC,GAAG,CAAC,CAAA;QACvB,MAAM,SAAS,GAAG,IAAI,CAAC,WAAW,EAAE,CAAA;QAEpC,cAAc;QACd,IAAI,CAAC,QAAQ,EAAE,CAAC;YACd,wDAAwD;YACxD,KAAK,MAAM,CAAC,GAAG,EAAE,GAAG,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,qBAAqB,CAAC,EAAE,CAAC;gBAC/D,2FAA2F;gBAC3F,IAAI,SAAS,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,GAAG,CAAC,IAAI,GAAG,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC;oBAC3E,QAAQ,GAAG,GAAG,CAAA;oBACd,MAAK;gBACP,CAAC;YACH,CAAC;YAED,mDAAmD;YACnD,IAAI,CAAC,QAAQ,EAAE,CAAC;gBACd,MAAM,cAAc,GAAG,SAAS,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,GAAG,CAAC,CAAA;gBAC3G,IAAI,cAAc,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBAC9B,KAAK,MAAM,GAAG,IAAI,aAAa,EAAE,CAAC;wBAChC,MAAM,aAAa,GAAG,GAAG,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,GAAG,CAAC,CAAA;wBAC5F,IAAI,cAAc,CAAC,QAAQ,CAAC,aAAa,CAAC,IAAI,aAAa,CAAC,QAAQ,CAAC,cAAc,CAAC,EAAE,CAAC;4BACrF,QAAQ,GAAG,GAAG,CAAA;4BACd,MAAK;wBACP,CAAC;oBACH,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;QAED,YAAY;QACZ,IAAI,CAAC,MAAM,EAAE,CAAC;YACZ,MAAM,UAAU,GAAG,aAAa,CAAC,SAAS,EAAE,UAAU,EAAE,CAAC,CAAC,CAAA;YAC1D,IAAI,UAAU,EAAE,CAAC;gBACf,MAAM,GAAG,UAAU,CAAA;YACrB,CAAC;iBAAM,CAAC;gBACN,KAAK,MAAM,CAAC,IAAI,UAAU,EAAE,CAAC;oBAC3B,MAAM,MAAM,GAAG,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,GAAG,CAAC,CAAA;oBACnC,IAAI,SAAS,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,SAAS,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,CAAC;wBACxD,MAAM,GAAG,CAAC,CAAA;wBACV,MAAK;oBACP,CAAC;oBACD,IAAI,SAAS,CAAC,MAAM,IAAI,MAAM,CAAC,MAAM,EAAE,CAAC;wBACtC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,SAAS,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;4BAC3D,MAAM,MAAM,GAAG,SAAS,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,CAAC,CAAA;4BACxD,MAAM,IAAI,GAAG,aAAa,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAA;4BAC1C,IAAI,IAAI,EAAE,CAAC;gCACT,MAAM,GAAG,CAAC,CAAA;gCACV,MAAK;4BACP,CAAC;wBACH,CAAC;oBACH,CAAC;oBACD,IAAI,MAAM;wBAAE,MAAK;gBACnB,CAAC;YACH,CAAC;QACH,CAAC;QAED,kBAAkB;QAClB,IAAI,CAAC,gBAAgB,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,cAAc,CAAC,IAAI,SAAS,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,EAAE,CAAC;YAC5F,IAAI,CAAC,GAAG,CAAC,CAAA;YACT,IAAI,CAAC,GAAG,CAAC,CAAA;YACT,IAAI,CAAC,GAAG,CAAC,CAAA;YACT,IAAI,KAAK,GAAG,KAAK,CAAA;YACjB,MAAM,MAAM,GAAG,SAAS,CAAC,KAAK,CAAC,WAAW,CAAC,CAAA;YAC3C,IAAI,MAAM,EAAE,CAAC;gBACX,CAAC,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAA;gBAC3B,KAAK,GAAG,IAAI,CAAA;YACd,CAAC;YACD,MAAM,MAAM,GAAG,SAAS,CAAC,KAAK,CAAC,WAAW,CAAC,CAAA;YAC3C,IAAI,MAAM,EAAE,CAAC;gBACX,CAAC,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAA;gBAC3B,KAAK,GAAG,IAAI,CAAA;YACd,CAAC;YACD,MAAM,MAAM,GAAG,SAAS,CAAC,KAAK,CAAC,WAAW,CAAC,CAAA;YAC3C,IAAI,MAAM,EAAE,CAAC;gBACX,CAAC,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAA;gBAC3B,KAAK,GAAG,IAAI,CAAA;YACd,CAAC;YACD,IAAI,KAAK,EAAE,CAAC;gBACV,gBAAgB,GAAG,CAAC,GAAG,KAAK,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,GAAG,EAAE,CAAA;YAClD,CAAC;QACH,CAAC;QAED,gBAAgB;QAChB,MAAM,WAAW,GAA2B;YAC1C,YAAY,EAAE,cAAc;YAC5B,cAAc,EAAE,cAAc;YAC9B,aAAa,EAAE,aAAa;SAC7B,CAAA;QACD,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,CAAA;QACrC,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;YACvB,IAAI,SAAS,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;gBAC5B,MAAM,OAAO,GAAG,SAAS,CAAC,KAAK,CAAC,cAAc,CAAC,CAAA;gBAC/C,IAAI,OAAO,EAAE,CAAC;oBACZ,MAAM,SAAS,GAAG,OAAO,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAA;oBAC/D,MAAM,MAAM,GAAG,QAAQ,CAAC,SAAS,CAAC,CAAA;oBAClC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC;wBACnB,MAAM,UAAU,GAAG,WAAW,CAAC,GAAG,CAAC,CAAA;wBACnC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,UAAU,CAAC,EAAE,CAAC;4BACjD,QAAQ,CAAC,IAAI,CAAC;gCACZ,UAAU,EAAE,kBAAkB;gCAC9B,IAAI,EAAE,UAAU;gCAChB,MAAM;gCACN,WAAW,EAAE,SAAS;6BACvB,CAAC,CAAA;wBACJ,CAAC;oBACH,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;QAED,6BAA6B;QAC7B,IAAI,cAAc,GAAG,KAAK,CAAA;QAC1B,IAAI,CAAC,kBAAkB,EAAE,CAAC;YACxB,IACE,SAAS,CAAC,QAAQ,CAAC,oBAAoB,CAAC;gBACxC,CAAC,SAAS,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,SAAS,CAAC,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,WAAW,CAAC,CAAC;gBACjG,CAAC,SAAS,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,SAAS,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,EAC1D,CAAC;gBACD,kBAAkB,GAAG,IAAI,CAAA;gBACzB,SAAQ;YACV,CAAC;QACH,CAAC;aAAM,CAAC;YACN,iEAAiE;YACjE,IACE,SAAS,CAAC,QAAQ,CAAC,YAAY,CAAC;gBAChC,SAAS,CAAC,QAAQ,CAAC,cAAc,CAAC;gBAClC,SAAS,CAAC,QAAQ,CAAC,iBAAiB,CAAC;gBACrC,SAAS,CAAC,QAAQ,CAAC,uBAAuB,CAAC,WAAW,EAAE,CAAC,EACzD,CAAC;gBACD,cAAc,GAAG,IAAI,CAAA;YACvB,CAAC;QACH,CAAC;QAED,IAAI,kBAAkB,EAAE,CAAC;YACvB,MAAM,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,KAAK,CAAC,CAAA;YACpC,IAAI,SAAS,GAAuB,IAAI,CAAA;YACxC,IAAI,QAAQ,GAAG,CAAC,CAAC,CAAA;YAEjB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACtC,MAAM,MAAM,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC,CAAA;gBAC5D,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC;oBAAE,SAAQ;gBAC/B,MAAM,KAAK,GAAG,aAAa,CAAC,MAAM,EAAE,OAAO,EAAE,CAAC,CAAC,CAAA;gBAC/C,IAAI,KAAK,EAAE,CAAC;oBACV,SAAS,GAAG,KAAK,CAAA;oBACjB,QAAQ,GAAG,CAAC,CAAA;oBACZ,MAAK;gBACP,CAAC;YACH,CAAC;YAED,IAAI,SAAS,EAAE,CAAC;gBACd,IAAI,OAAO,GAAa,EAAE,CAAA;gBAC1B,KAAK,IAAI,CAAC,GAAG,QAAQ,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;oBACjD,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC;wBAAE,SAAQ;oBACpC,MAAM,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAA;oBAC5C,IAAI,GAAG,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;wBACnB,MAAM,GAAG,GAAG,UAAU,CAAC,GAAG,CAAC,CAAA;wBAC3B,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC;4BAAE,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;oBACpC,CAAC;gBACH,CAAC;gBAED,2DAA2D;gBAC3D,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,IAAI,GAAG,GAAG,CAAC,EAAE,CAAC;oBAClC,MAAM,QAAQ,GAAG,KAAK,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,WAAW,EAAE,CAAA;oBAC7C,iFAAiF;oBACjF,iEAAiE;oBACjE,IAAI,cAAc,GAAG,KAAK,CAAA;oBAC1B,MAAM,SAAS,GAAG,QAAQ,CAAC,KAAK,CAAC,KAAK,CAAC,CAAA;oBACvC,KAAK,MAAM,CAAC,IAAI,SAAS,EAAE,CAAC;wBAC1B,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC,CAAA;wBAChD,IAAI,CAAC,CAAC,MAAM,IAAI,CAAC,IAAI,aAAa,CAAC,CAAC,EAAE,OAAO,EAAE,CAAC,CAAC,EAAE,CAAC;4BAClD,cAAc,GAAG,IAAI,CAAA;4BACrB,MAAK;wBACP,CAAC;oBACH,CAAC;oBAED,IAAI,CAAC,cAAc,EAAE,CAAC;wBACpB,4CAA4C;wBAC5C,MAAM,WAAW,GAAa,EAAE,CAAA;wBAChC,KAAK,MAAM,CAAC,IAAI,SAAS,EAAE,CAAC;4BAC1B,IAAI,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC;gCAAE,SAAQ;4BAC7B,MAAM,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAA;4BACrC,IAAI,GAAG,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gCACnB,MAAM,GAAG,GAAG,UAAU,CAAC,GAAG,CAAC,CAAA;gCAC3B,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC;oCAAE,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;4BACxC,CAAC;wBACH,CAAC;wBACD,IAAI,WAAW,CAAC,MAAM,IAAI,CAAC,EAAE,CAAC;4BAC5B,OAAO,GAAG,WAAW,CAAA;wBACvB,CAAC;oBACH,CAAC;gBACH,CAAC;gBAED,IAAI,OAAO,CAAC,MAAM,IAAI,CAAC,EAAE,CAAC;oBACxB,MAAM,QAAQ,GAAG,OAAO,CAAC,CAAC,CAAC,CAAA;oBAC3B,QAAQ,CAAC,IAAI,CAAC;wBACZ,UAAU,EAAE,oBAAoB;wBAChC,GAAG,EAAE,SAAS;wBACd,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC;qBAC5B,CAAC,CAAA;gBACJ,CAAC;YACH,CAAC;YAED,IAAI,cAAc,EAAE,CAAC;gBACnB,kBAAkB,GAAG,KAAK,CAAA;YAC5B,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO;QACL,UAAU,EAAE,wBAAwB;QACpC,QAAQ;QACR,MAAM;QACN,gBAAgB;QAChB,QAAQ,EAAE,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,SAAS;QACpD,QAAQ;KACT,CAAA;AACH,CAAC,CAAA"}

package/dist/parsers/shipRock.d.ts

@@ -0,0 +1,2 @@
+import { ShipRockInput } from '@regolithco/common';
+export declare const parseShipRock: (ocrText: string) => ShipRockInput;

package/dist/parsers/shipRock.js

@@ -0,0 +1,156 @@
+import { ShipOreEnum, RockStateEnum, AsteroidTypeEnum, DepositTypeEnum, } from '@regolithco/common';
+import { cleanLine, levenshtein } from './utils.js';
+export const parseShipRock = (ocrText) => {
+    // Split into lines to preserve spatial context
+    const lines = ocrText
+        .split('\n')
+        .map(cleanLine)
+        .filter((l) => l.length > 0);
+    // Default values
+    let mass = 0;
+    let inst;
+    let res;
+    let rockType;
+    const ores = [];
+    const allOres = Object.values(ShipOreEnum);
+    const asteroidMap = {
+        'M-TYPE': AsteroidTypeEnum.Mtype,
+        'C-TYPE': AsteroidTypeEnum.Ctype,
+        'S-TYPE': AsteroidTypeEnum.Stype,
+        'E-TYPE': AsteroidTypeEnum.Etype,
+        'E TYPE': AsteroidTypeEnum.Etype,
+        'I-TYPE': AsteroidTypeEnum.Itype,
+        'P-TYPE': AsteroidTypeEnum.Ptype,
+        'Q-TYPE': AsteroidTypeEnum.Qtype,
+    };
+    // Iterate lines to find tokens
+    for (const line of lines) {
+        const lineUpper = line.toUpperCase();
+        // 1. Mass
+        if (!mass) {
+            // Strict Mass: "MASS: 1234"
+            const massMatch = line.match(/Mass\s*[:=\s]\s*([\d,]+(?:\.\d+)?)/i);
+            if (massMatch) {
+                const val = parseFloat(massMatch[1].replace(/,/g, ''));
+                if (!isNaN(val))
+                    mass = val;
+            }
+        }
+        // 2. Instability
+        if (inst === undefined) {
+            const match = line.match(/(?:Inst|ABIT|STAB)[^\d]*([\d.]+)/i);
+            if (match)
+                inst = parseFloat(match[1]);
+        }
+        // 3. Resistance
+        if (res === undefined) {
+            // RESISTANCE: 10%
+            const match = line.match(/(?:Res|SISTAN|RESIST)[^\d]*([\d.]+)/i);
+            if (match) {
+                let val = parseFloat(match[1]);
+                if (line.includes('%') && val > 1)
+                    val /= 100;
+                res = val;
+            }
+        }
+        // 4. Rock Type
+        if (!rockType) {
+            for (const [key, value] of Object.entries(asteroidMap)) {
+                // Flexible match
+                const cleanKey = key.replace(/[- ]/g, '[\\- _\\.]?');
+                const regex = new RegExp(cleanKey, 'i');
+                if (regex.test(line)) {
+                    rockType = value;
+                    break;
+                }
+            }
+            // Deposit types if not asteroid
+            if (!rockType) {
+                const depositTypes = Object.values(DepositTypeEnum);
+                // Only strict match on line for deposit types to avoid noise
+                for (const dt of depositTypes) {
+                    if (lineUpper.includes(dt.toUpperCase())) {
+                        rockType = dt;
+                        break;
+                    }
+                }
+            }
+        }
+        // 5. Ores
+        // Refined logic: Iterate words in line. Identify Ores. Look for adjacent %.
+        const words = lineUpper.split(/\s+/);
+        for (let i = 0; i < words.length; i++) {
+            const w = words[i].replace(/[^A-Z0-9%.-]/g, ''); // Clean word
+            if (w.length < 3)
+                continue;
+            // Check if w is an Ore (Best match)
+            let bestDist = Infinity;
+            let bestOre = null;
+            for (const ore of allOres) {
+                let dist = levenshtein(w, ore.toUpperCase());
+                // Special cases
+                if (w === 'INERT') {
+                    if (ore === ShipOreEnum.Inertmaterial)
+                        dist = 0;
+                }
+                else if (w === 'MATERIAL' || w === 'MATERIALS') {
+                    // Context check: previous word INERT?
+                    if (i > 0 && words[i - 1].toUpperCase().includes('INERT') && ore === ShipOreEnum.Inertmaterial) {
+                        dist = 0;
+                    }
+                }
+                let threshold = 1;
+                if (w.length <= 3)
+                    threshold = 0;
+                if (dist <= threshold && dist < bestDist) {
+                    bestDist = dist;
+                    bestOre = ore;
+                }
+            }
+            if (bestOre) {
+                // We found an ore token. Look for adjacent percentage (i-1 or i+1)
+                let percentVal = 0;
+                // Check Prev
+                if (i > 0) {
+                    const prev = words[i - 1];
+                    // Allow bare number ONLY if it is a float (has decimal) or has %
+                    // This avoids matching "100" (Power) or "70" (Level) as percentage.
+                    if (prev.includes('%') || prev.includes('.')) {
+                        const val = parseFloat(prev);
+                        if (!isNaN(val))
+                            percentVal = val;
+                    }
+                }
+                // Check Next (if not found in prev)
+                if (percentVal === 0 && i < words.length - 1) {
+                    const next = words[i + 1];
+                    // Next MUST have % to avoid confusion with SCU volume
+                    if (next.includes('%')) {
+                        const val = parseFloat(next);
+                        if (!isNaN(val))
+                            percentVal = val;
+                    }
+                }
+                if (percentVal > 0) {
+                    if (percentVal > 1)
+                        percentVal /= 100;
+                    if (!ores.find((o) => o.ore === bestOre)) {
+                        ores.push({
+                            ore: bestOre,
+                            percent: percentVal,
+                        });
+                    }
+                }
+            }
+        }
+    }
+    return {
+        mass,
+        inst,
+        res,
+        rockType,
+        ores,
+        state: RockStateEnum.Ready,
+    };
+};
+//# sourceMappingURL=shipRock.js.map

package/dist/parsers/shipRock.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"shipRock.js","sourceRoot":"","sources":["../../src/parsers/shipRock.ts"],"names":[],"mappings":"AAAA,OAAO,EAIL,WAAW,EACX,aAAa,EACb,gBAAgB,EAChB,eAAe,GAChB,MAAM,oBAAoB,CAAA;AAC3B,OAAO,EAAE,SAAS,EAAE,WAAW,EAAE,MAAM,YAAY,CAAA;AAEnD,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,OAAe,EAAiB,EAAE;IAC9D,+CAA+C;IAC/C,MAAM,KAAK,GAAG,OAAO;SAClB,KAAK,CAAC,IAAI,CAAC;SACX,GAAG,CAAC,SAAS,CAAC;SACd,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAA;IAE9B,iBAAiB;IACjB,IAAI,IAAI,GAAG,CAAC,CAAA;IACZ,IAAI,IAAwB,CAAA;IAC5B,IAAI,GAAuB,CAAA;IAC3B,IAAI,QAA8B,CAAA;IAClC,MAAM,IAAI,GAAuB,EAAE,CAAA;IAEnC,MAAM,OAAO,GAAG,MAAM,CAAC,MAAM,CAAC,WAAW,CAAC,CAAA;IAC1C,MAAM,WAAW,GAAqC;QACpD,QAAQ,EAAE,gBAAgB,CAAC,KAAK;QAChC,QAAQ,EAAE,gBAAgB,CAAC,KAAK;QAChC,QAAQ,EAAE,gBAAgB,CAAC,KAAK;QAChC,QAAQ,EAAE,gBAAgB,CAAC,KAAK;QAChC,QAAQ,EAAE,gBAAgB,CAAC,KAAK;QAChC,QAAQ,EAAE,gBAAgB,CAAC,KAAK;QAChC,QAAQ,EAAE,gBAAgB,CAAC,KAAK;QAChC,QAAQ,EAAE,gBAAgB,CAAC,KAAK;KACjC,CAAA;IAED,+BAA+B;IAC/B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,SAAS,GAAG,IAAI,CAAC,WAAW,EAAE,CAAA;QAEpC,UAAU;QACV,IAAI,CAAC,IAAI,EAAE,CAAC;YACV,4BAA4B;YAC5B,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,qCAAqC,CAAC,CAAA;YACnE,IAAI,SAAS,EAAE,CAAC;gBACd,MAAM,GAAG,GAAG,UAAU,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,CAAA;gBACtD,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC;oBAAE,IAAI,GAAG,GAAG,CAAA;YAC7B,CAAC;QACH,CAAC;QAED,iBAAiB;QACjB,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;YACvB,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,mCAAmC,CAAC,CAAA;YAC7D,IAAI,KAAK;gBAAE,IAAI,GAAG,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAA;QACxC,CAAC;QAED,gBAAgB;QAChB,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;YACtB,kBAAkB;YAClB,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,sCAAsC,CAAC,CAAA;YAChE,IAAI,KAAK,EAAE,CAAC;gBACV,IAAI,GAAG,GAAG,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAA;gBAC9B,IAAI,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,GAAG,GAAG,CAAC;oBAAE,GAAG,IAAI,GAAG,CAAA;gBAC7C,GAAG,GAAG,GAAG,CAAA;YACX,CAAC;QACH,CAAC;QAED,eAAe;QACf,IAAI,CAAC,QAAQ,EAAE,CAAC;YACd,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,WAAW,CAAC,EAAE,CAAC;gBACvD,iBAAiB;gBACjB,MAAM,QAAQ,GAAG,GAAG,CAAC,OAAO,CAAC,OAAO,EAAE,aAAa,CAAC,CAAA;gBACpD,MAAM,KAAK,GAAG,IAAI,MAAM,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAA;gBACvC,IAAI,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;oBACrB,QAAQ,GAAG,KAAK,CAAA;oBAChB,MAAK;gBACP,CAAC;YACH,CAAC;YACD,gCAAgC;YAChC,IAAI,CAAC,QAAQ,EAAE,CAAC;gBACd,MAAM,YAAY,GAAG,MAAM,CAAC,MAAM,CAAC,eAAe,CAAC,CAAA;gBACnD,6DAA6D;gBAC7D,KAAK,MAAM,EAAE,IAAI,YAAY,EAAE,CAAC;oBAC9B,IAAI,SAAS,CAAC,QAAQ,CAAC,EAAE,CAAC,WAAW,EAAE,CAAC,EAAE,CAAC;wBACzC,QAAQ,GAAG,EAAqB,CAAA;wBAChC,MAAK;oBACP,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;QAED,UAAU;QACV,4EAA4E;QAC5E,MAAM,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,KAAK,CAAC,CAAA;QAEpC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACtC,MAAM,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,eAAe,EAAE,EAAE,CAAC,CAAA,CAAC,aAAa;YAC7D,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC;gBAAE,SAAQ;YAE1B,oCAAoC;YACpC,IAAI,QAAQ,GAAG,QAAQ,CAAA;YACvB,IAAI,OAAO,GAAuB,IAAI,CAAA;YAEtC,KAAK,MAAM,GAAG,IAAI,OAAO,EAAE,CAAC;gBAC1B,IAAI,IAAI,GAAG,WAAW,CAAC,CAAC,EAAE,GAAG,CAAC,WAAW,EAAE,CAAC,CAAA;gBAC5C,gBAAgB;gBAChB,IAAI,CAAC,KAAK,OAAO,EAAE,CAAC;oBAClB,IAAI,GAAG,KAAK,WAAW,CAAC,aAAa;wBAAE,IAAI,GAAG,CAAC,CAAA;gBACjD,CAAC;qBAAM,IAAI,CAAC,KAAK,UAAU,IAAI,CAAC,KAAK,WAAW,EAAE,CAAC;oBACjD,sCAAsC;oBACtC,IAAI,CAAC,GAAG,CAAC,IAAI,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,GAAG,KAAK,WAAW,CAAC,aAAa,EAAE,CAAC;wBAC/F,IAAI,GAAG,CAAC,CAAA;oBACV,CAAC;gBACH,CAAC;gBAED,IAAI,SAAS,GAAG,CAAC,CAAA;gBACjB,IAAI,CAAC,CAAC,MAAM,IAAI,CAAC;oBAAE,SAAS,GAAG,CAAC,CAAA;gBAEhC,IAAI,IAAI,IAAI,SAAS,IAAI,IAAI,GAAG,QAAQ,EAAE,CAAC;oBACzC,QAAQ,GAAG,IAAI,CAAA;oBACf,OAAO,GAAG,GAAG,CAAA;gBACf,CAAC;YACH,CAAC;YAED,IAAI,OAAO,EAAE,CAAC;gBACZ,mEAAmE;gBACnE,IAAI,UAAU,GAAG,CAAC,CAAA;gBAElB,aAAa;gBACb,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;oBACV,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;oBACzB,iEAAiE;oBACjE,oEAAoE;oBACpE,IAAI,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;wBAC7C,MAAM,GAAG,GAAG,UAAU,CAAC,IAAI,CAAC,CAAA;wBAC5B,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC;4BAAE,UAAU,GAAG,GAAG,CAAA;oBACnC,CAAC;gBACH,CAAC;gBAED,oCAAoC;gBACpC,IAAI,UAAU,KAAK,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBAC7C,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;oBACzB,sDAAsD;oBACtD,IAAI,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;wBACvB,MAAM,GAAG,GAAG,UAAU,CAAC,IAAI,CAAC,CAAA;wBAC5B,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC;4BAAE,UAAU,GAAG,GAAG,CAAA;oBACnC,CAAC;gBACH,CAAC;gBAED,IAAI,UAAU,GAAG,CAAC,EAAE,CAAC;oBACnB,IAAI,UAAU,GAAG,CAAC;wBAAE,UAAU,IAAI,GAAG,CAAA;oBAErC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,GAAG,KAAK,OAAO,CAAC,EAAE,CAAC;wBACzC,IAAI,CAAC,IAAI,CAAC;4BACR,GAAG,EAAE,OAAO;4BACZ,OAAO,EAAE,UAAU;yBACpB,CAAC,CAAA;oBACJ,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO;QACL,IAAI;QACJ,IAAI;QACJ,GAAG;QACH,QAAQ;QACR,IAAI;QACJ,KAAK,EAAE,aAAa,CAAC,KAAK;KAC3B,CAAA;AACH,CAAC,CAAA"}

package/dist/parsers/utils.d.ts

@@ -0,0 +1,3 @@
+export declare const cleanLine: (text: string) => string;
+export declare const levenshtein: (a: string, b: string) => number;
+export declare const findBestMatch: <T extends string>(text: string, candidates: T[], threshold?: number) => T | null;

package/dist/parsers/utils.js

@@ -0,0 +1,32 @@
+export const cleanLine = (text) => {
+    return text.trim();
+};
+export const levenshtein = (a, b) => {
+    const matrix = Array.from({ length: a.length + 1 }, () => Array(b.length + 1).fill(0));
+    for (let i = 0; i <= a.length; i++)
+        matrix[i][0] = i;
+    for (let j = 0; j <= b.length; j++)
+        matrix[0][j] = j;
+    for (let i = 1; i <= a.length; i++) {
+        for (let j = 1; j <= b.length; j++) {
+            const cost = a[i - 1] === b[j - 1] ? 0 : 1;
+            matrix[i][j] = Math.min(matrix[i - 1][j] + 1, matrix[i][j - 1] + 1, matrix[i - 1][j - 1] + cost);
+        }
+    }
+    return matrix[a.length][b.length];
+};
+export const findBestMatch = (text, candidates, threshold = 2) => {
+    let bestDist = Infinity;
+    let bestMatch = null;
+    const upperText = text.toUpperCase();
+    for (const cand of candidates) {
+        const upperCand = cand.toUpperCase();
+        const dist = levenshtein(upperText, upperCand);
+        if (dist < bestDist && dist <= threshold) {
+            bestDist = dist;
+            bestMatch = cand;
+        }
+    }
+    return bestMatch;
+};
+//# sourceMappingURL=utils.js.map

package/dist/parsers/utils.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"utils.js","sourceRoot":"","sources":["../../src/parsers/utils.ts"],"names":[],"mappings":"AAAA,MAAM,CAAC,MAAM,SAAS,GAAG,CAAC,IAAY,EAAU,EAAE;IAChD,OAAO,IAAI,CAAC,IAAI,EAAE,CAAA;AACpB,CAAC,CAAA;AAED,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,CAAS,EAAE,CAAS,EAAU,EAAE;IAC1D,MAAM,MAAM,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,MAAM,GAAG,CAAC,EAAE,EAAE,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAA;IACtF,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE;QAAE,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;IACpD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE;QAAE,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;IAEpD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACnC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACnC,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;YAC1C,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,CAAA;QAClG,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAA;AACnC,CAAC,CAAA;AAED,MAAM,CAAC,MAAM,aAAa,GAAG,CAAmB,IAAY,EAAE,UAAe,EAAE,YAAoB,CAAC,EAAY,EAAE;IAChH,IAAI,QAAQ,GAAG,QAAQ,CAAA;IACvB,IAAI,SAAS,GAAa,IAAI,CAAA;IAC9B,MAAM,SAAS,GAAG,IAAI,CAAC,WAAW,EAAE,CAAA;IAEpC,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;QAC9B,MAAM,SAAS,GAAG,IAAI,CAAC,WAAW,EAAE,CAAA;QACpC,MAAM,IAAI,GAAG,WAAW,CAAC,SAAS,EAAE,SAAS,CAAC,CAAA;QAC9C,IAAI,IAAI,GAAG,QAAQ,IAAI,IAAI,IAAI,SAAS,EAAE,CAAC;YACzC,QAAQ,GAAG,IAAI,CAAA;YACf,SAAS,GAAG,IAAI,CAAA;QAClB,CAAC;IACH,CAAC;IACD,OAAO,SAAS,CAAA;AAClB,CAAC,CAAA"}

package/package.json

@@ -0,0 +1,65 @@
+{
+  "name": "@regolithco/ocr",
+  "version": "1.0.0",
+  "description": "OCR utilities using PaddleOCR",
+  "type": "module",
+  "main": "dist/node.js",
+  "browser": "dist/browser.js",
+  "types": "dist/node.d.ts",
+  "publishConfig": {
+    "access": "public"
+  },
+  "engines": {
+    "node": ">=22 <24"
+  },  
+  "files": [
+    "dist"
+  ],
+  "scripts": {
+    "build": "tsc",
+    "prepublishOnly": "yarn run build",
+    "debug": "tsx scripts/debug_paddle.ts",
+    "debug:choose": "tsx scripts/debug_paddle.ts --choose",
+    "test": "node --experimental-vm-modules node_modules/jest/bin/jest.js"
+  },
+  "keywords": [
+    "ocr",
+    "paddle",
+    "regolith"
+  ],
+  "author": "",
+  "license": "ISC",
+  "dependencies": {
+    "@gutenye/ocr-browser": "^1.4.8",
+    "@gutenye/ocr-node": "^1.4.8",
+    "@regolithco/common": "^1.4.7",
+    "graphql": "^16.12.0",
+    "loglevel": "^1.9.2",
+    "onnxruntime-node": "^1.23.2"
+  },
+  "devDependencies": {
+    "@clack/prompts": "^0.11.0",
+    "@faker-js/faker": "8",
+    "@types/jest": "^30.0.0",
+    "@types/lodash": "^4.17.20",
+    "@types/node": "^24.5.2",
+    "@types/uuid": "9",
+    "@typescript-eslint/eslint-plugin": "^8.44.1",
+    "@typescript-eslint/parser": "^8.44.1",
+    "eslint": "^9.36.0",
+    "eslint-config-prettier": "^10.1.8",
+    "eslint-config-standard": "^17.1.0",
+    "eslint-plugin-import": "^2.32.0",
+    "eslint-plugin-n": "^17.23.1",
+    "eslint-plugin-node": "^11.1.0",
+    "eslint-plugin-prettier": "^5.5.4",
+    "eslint-plugin-promise": "^7.2.1",
+    "jest": "^30.2.0",
+    "prettier": "^3.6.2",
+    "ts-jest": "^29.4.4",
+    "ts-node": "^10.9.2",
+    "tsx": "^4.21.0",
+    "typescript": "^5.9.2",
+    "uuid": "^13.0.0"
+  }
+}