openrxiv-cli 0.0.2

package/dist/commands/list.js

@@ -0,0 +1,18 @@
+import { Command } from 'commander';
+import { listBucketContent } from '../aws/bucket-explorer.js';
+import { getDefaultServer } from '../utils/index.js';
+export const listCommand = new Command('list')
+    .description('List available content in the bioRxiv or medRxiv S3 bucket')
+    .option('-m, --month <month>', 'Filter by specific month (e.g., "2024-01")')
+    .option('-b, --batch <batch>', 'Filter by specific batch (e.g., "Batch_01")')
+    .option('-l, --limit <number>', 'Limit the number of results', '50')
+    .option('-s, --server <server>', 'Server to use: "biorxiv" or "medrxiv"', getDefaultServer())
+    .action(async (options) => {
+    try {
+        await listBucketContent(options);
+    }
+    catch (error) {
+        console.error('Error listing content:', error);
+        process.exit(1);
+    }
+});

package/dist/commands/summary.d.ts

@@ -0,0 +1,3 @@
+import { Command } from 'commander';
+export declare const summaryCommand: Command;
+//# sourceMappingURL=summary.d.ts.map

package/dist/commands/summary.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"summary.d.ts","sourceRoot":"","sources":["../../src/commands/summary.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAOpC,eAAO,MAAM,cAAc,SAyGvB,CAAC"}

package/dist/commands/summary.js

@@ -0,0 +1,249 @@
+import { Command } from 'commander';
+import chalk from 'chalk';
+import boxen from 'boxen';
+import { createOpenRxivApiClient, getServerFromDOI } from '../api/api-client.js';
+import { parseBiorxivURL } from 'openrxiv-utils';
+import { getDefaultServer } from '../utils/index.js';
+export const summaryCommand = new Command('summary')
+    .description('Get a summary of a bioRxiv preprint from a URL or DOI')
+    .argument('<url-or-doi>', 'bioRxiv URL or DOI to summarize')
+    .option('-m, --more', 'Show additional details and full abstract')
+    .option('-s, --server <server>', 'Specify server (biorxiv or medrxiv)', getDefaultServer())
+    .action(async (urlOrDoi, options) => {
+    var _a, _b;
+    try {
+        console.log(chalk.blue.bold(`🔬 ${(_a = options.server) !== null && _a !== void 0 ? _a : ''} Preprint Summary`));
+        console.log(chalk.blue('================================\n'));
+        // Parse the input (could be URL or DOI)
+        let doi;
+        let parsedUrl = null;
+        if (urlOrDoi.includes('biorxiv.org') ||
+            urlOrDoi.includes('medrxiv.org') ||
+            urlOrDoi.includes('doi.org')) {
+            // It's a URL
+            parsedUrl = parseBiorxivURL(urlOrDoi);
+            if (!parsedUrl) {
+                console.log(chalk.red('❌ Invalid bioRxiv URL'));
+                process.exit(1);
+            }
+            doi = parsedUrl.doi;
+            console.log(chalk.cyan(`📝 URL: ${urlOrDoi}`));
+        }
+        else {
+            // It's a DOI
+            doi = urlOrDoi;
+            console.log(chalk.cyan(`🔍 DOI: ${doi}`));
+        }
+        console.log('');
+        // Determine server if possible
+        let server = (_b = options.server) !== null && _b !== void 0 ? _b : getServerFromDOI(urlOrDoi);
+        console.log(chalk.blue(`🌐 Server: ${server}`));
+        console.log('');
+        // Create API client
+        const apiClient = createOpenRxivApiClient({
+            server,
+            format: 'json',
+            timeout: 15000,
+        });
+        // Show loading indicator
+        console.log(chalk.yellow('⏳ Fetching preprint information...'));
+        console.log('');
+        // Get content details
+        let contentDetail = await apiClient.getContentDetail(doi);
+        let fallbackServer = null;
+        // If not found on bioRxiv and we're not already on medrxiv, try medrxiv as fallback
+        if (!contentDetail && server === 'biorxiv') {
+            console.log(chalk.yellow('⚠️  Paper not found on bioRxiv, trying medRxiv...'));
+            fallbackServer = 'medrxiv';
+            const medrxivApiClient = createOpenRxivApiClient({
+                server: 'medrxiv',
+                format: 'json',
+                timeout: 15000,
+            });
+            contentDetail = await medrxivApiClient.getContentDetail(doi);
+            if (contentDetail) {
+                console.log(chalk.green('✅ Found paper on medRxiv!'));
+                server = 'medrxiv'; // Update server for display
+                contentDetail.server = 'medrxiv'; // Ensure the content detail has the correct server
+            }
+        }
+        if (!contentDetail) {
+            console.log(chalk.red('❌ No content found for this DOI on either bioRxiv or medRxiv'));
+            console.log(chalk.yellow("💡 This might be a new preprint that hasn't been indexed yet"));
+            process.exit(1);
+        }
+        // Get all versions
+        let allVersions = await apiClient.getAllVersions(doi);
+        // If we used fallback, get versions from the fallback server
+        if (fallbackServer && contentDetail) {
+            const fallbackApiClient = createOpenRxivApiClient({
+                server: fallbackServer,
+                format: 'json',
+                timeout: 15000,
+            });
+            allVersions = await fallbackApiClient.getAllVersions(doi);
+        }
+        // Display summary
+        const isVerbose = options.more === true;
+        displaySummary(contentDetail, allVersions, isVerbose);
+    }
+    catch (error) {
+        console.error(chalk.red('❌ Error:'), error instanceof Error ? error.message : 'Unknown error');
+        process.exit(1);
+    }
+});
+function displaySummary(contentDetail, allVersions, verbose = false) {
+    // Title in a prominent box
+    const titleBox = boxen(chalk.green.bold.underline(contentDetail.title), {
+        padding: 1,
+        margin: 1,
+        borderStyle: 'double',
+        borderColor: 'green',
+        backgroundColor: 'black',
+        textAlignment: 'left',
+        ...(verbose ? {} : { width: 80 }),
+    });
+    console.log(titleBox);
+    // Basic info
+    const basicInfo = [
+        `${chalk.cyan('DOI:')} ${contentDetail.doi}`,
+        `${chalk.cyan('Server:')} ${contentDetail.server}`,
+        `${chalk.cyan('Category:')} ${chalk.yellow(contentDetail.category)}`,
+        `${chalk.cyan('License:')} ${chalk.yellow(contentDetail.license)}`,
+        `${chalk.cyan('Type:')} ${chalk.yellow(contentDetail.type)}`,
+        `${chalk.cyan('Published:')} ${contentDetail.published === 'NA'
+            ? chalk.gray('Not published')
+            : chalk.green(contentDetail.published)}`,
+        `${chalk.cyan('Total versions:')} ${allVersions ? allVersions.length : 0}`,
+        ...(verbose
+            ? [
+                `${chalk.cyan('Date:')} ${contentDetail.date}`,
+                `${chalk.cyan('Version:')} ${contentDetail.version}`,
+                ...(contentDetail.jatsxml
+                    ? [`${chalk.cyan('JATS XML:')} ${chalk.underline.blue(contentDetail.jatsxml)}`]
+                    : []),
+            ]
+            : []),
+    ].join('\n');
+    const basicInfoBox = boxen(basicInfo, {
+        padding: 1,
+        margin: 1,
+        borderStyle: 'round',
+        borderColor: 'blue',
+        title: chalk.blue.bold('📋 Basic Information'),
+        titleAlignment: 'left',
+        textAlignment: 'left',
+        ...(verbose ? {} : { width: 80 }),
+    });
+    console.log(basicInfoBox);
+    // Authors
+    const authorsInfo = [
+        `${chalk.cyan('Authors:')} ${contentDetail.authors}`,
+        ...(contentDetail.author_corresponding
+            ? [`${chalk.cyan('Corresponding:')} ${chalk.green(contentDetail.author_corresponding)}`]
+            : []),
+        ...(contentDetail.author_corresponding_institution
+            ? [
+                `${chalk.cyan('Institution:')} ${chalk.gray(contentDetail.author_corresponding_institution)}`,
+            ]
+            : []),
+    ].join('\n');
+    const authorsBox = boxen(authorsInfo, {
+        padding: 1,
+        margin: 1,
+        borderStyle: 'round',
+        borderColor: 'cyan',
+        title: chalk.cyan.bold('👥 Authors'),
+        titleAlignment: 'left',
+        textAlignment: 'left',
+        ...(verbose ? {} : { width: 80 }),
+    });
+    console.log(authorsBox);
+    // Abstract
+    if (contentDetail.abstract) {
+        const abstractBox = boxen(contentDetail.abstract, {
+            padding: 1,
+            margin: 1,
+            borderStyle: 'round',
+            borderColor: 'yellow',
+            title: chalk.yellow.bold('📖 Abstract'),
+            titleAlignment: 'left',
+            textAlignment: 'left',
+            ...(verbose ? {} : { width: 80 }),
+        });
+        console.log(abstractBox);
+    }
+    // Funding information
+    if (contentDetail.funding && contentDetail.funding.length > 0) {
+        const fundingInfo = contentDetail.funding
+            .map((fund, index) => {
+            let fundText = `${chalk.cyan(index + 1)}. ${chalk.green(fund.name)}`;
+            if (fund.id) {
+                fundText += `\n   ${chalk.gray('ID:')} ${fund.id} (${fund['id-type']})`;
+            }
+            if (fund.award) {
+                fundText += `\n   ${chalk.gray('Award:')} ${fund.award}`;
+            }
+            return fundText;
+        })
+            .join('\n\n');
+        const fundingBox = boxen(fundingInfo, {
+            padding: 1,
+            margin: 1,
+            borderStyle: 'round',
+            borderColor: 'magenta',
+            title: chalk.magenta.bold('💰 Funding'),
+            titleAlignment: 'left',
+            textAlignment: 'left',
+            ...(verbose ? {} : { width: 80 }),
+        });
+        console.log(fundingBox);
+    }
+    // Versions - only show when --more is provided
+    if (verbose && allVersions && allVersions.length > 0) {
+        const versionsInfo = [
+            ...allVersions.map((version, index) => {
+                let versionText = `${chalk.cyan(`v${version.version}`)} (${version.date})`;
+                versionText += `\n   ${chalk.gray('Type:')} ${version.type}`;
+                versionText += `\n   ${chalk.gray('Title:')} ${version.title.substring(0, 60)}...`;
+                if (version.jatsxml) {
+                    versionText += `\n   ${chalk.gray('JATS XML:')} ${chalk.underline.blue(version.jatsxml)}`;
+                }
+                return versionText;
+            }),
+        ].join('\n\n');
+        const versionsBox = boxen(versionsInfo, {
+            padding: 1,
+            margin: 1,
+            borderStyle: 'round',
+            borderColor: 'green',
+            title: chalk.green.bold('🔄 Versions'),
+            titleAlignment: 'left',
+            textAlignment: 'left',
+            ...(verbose ? {} : { width: 80 }),
+        });
+        console.log(versionsBox);
+    }
+    // Footer
+    const baseUrl = contentDetail.server === 'medrxiv'
+        ? `https://www.medrxiv.org/content/${contentDetail.doi}`
+        : `https://www.biorxiv.org/content/${contentDetail.doi}`;
+    const footerInfo = [
+        `💡 View online: ${chalk.underline.blue(baseUrl)}`,
+        ...(allVersions && allVersions.length > 1 && !verbose
+            ? [
+                `📚 This preprint has ${allVersions.length} versions. Use --more to see additional details.`,
+            ]
+            : []),
+    ].join('\n');
+    const footerBox = boxen(footerInfo, {
+        padding: 1,
+        margin: 1,
+        borderStyle: 'round',
+        borderColor: 'gray',
+        backgroundColor: 'black',
+        textAlignment: 'left',
+        ...(verbose ? {} : { width: 80 }),
+    });
+    console.log(footerBox);
+}

package/dist/index.d.ts

@@ -0,0 +1,7 @@
+import { Command } from 'commander';
+export * from './commands/index.js';
+export { default as version } from './version.js';
+export { getCliName } from './utils/index.js';
+export { setGlobalRequesterPays } from './aws/config.js';
+export declare function createCLI(): Command;
+//# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAYpC,cAAc,qBAAqB,CAAC;AACpC,OAAO,EAAE,OAAO,IAAI,OAAO,EAAE,MAAM,cAAc,CAAC;AAClD,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAC9C,OAAO,EAAE,sBAAsB,EAAE,MAAM,iBAAiB,CAAC;AAEzD,wBAAgB,SAAS,IAAI,OAAO,CAqCnC"}

package/dist/index.js

@@ -0,0 +1,35 @@
+import { Command } from 'commander';
+import { listCommand, downloadCommand, summaryCommand, monthInfoCommand, batchProcessCommand, } from './commands/index.js';
+import { getCliName } from './utils/index.js';
+import { setGlobalRequesterPays } from './aws/config.js';
+import version from './version.js';
+export * from './commands/index.js';
+export { default as version } from './version.js';
+export { getCliName } from './utils/index.js';
+export { setGlobalRequesterPays } from './aws/config.js';
+export function createCLI() {
+    const cliName = getCliName();
+    const program = new Command();
+    program
+        .name(cliName)
+        .description(`CLI tool to download bioRxiv/medRxiv MECA files from AWS S3 for text and data mining`)
+        .version(`v${version}`, '-v, --version', `Print the current version of the ${cliName} CLI`);
+    // Add commands
+    program.addCommand(listCommand);
+    program.addCommand(downloadCommand);
+    program.addCommand(summaryCommand);
+    program.addCommand(monthInfoCommand);
+    program.addCommand(batchProcessCommand);
+    // Global options
+    program.option('-d, --debug', 'Enable debug mode');
+    program.option('--requester-pays', 'Enable requester pays for local development (required when not on EC2 with IAM role)');
+    // Parse command line arguments
+    program.parse();
+    // Set global requester pays based on command line option
+    const options = program.opts();
+    if (options.requesterPays) {
+        setGlobalRequesterPays(true);
+        console.log('Requester pays enabled - you will be charged for S3 requests');
+    }
+    return program;
+}

package/dist/utils/batches.d.ts

@@ -0,0 +1,9 @@
+/**
+ * Parse batch input to support ranges like "1-10" or "batch-1,batch-2"
+ */
+export declare function parseBatchInput(batchInput: string): string[];
+/**
+ * Validate batch format
+ */
+export declare function validateBatchFormat(batch: string): boolean;
+//# sourceMappingURL=batches.d.ts.map

package/dist/utils/batches.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"batches.d.ts","sourceRoot":"","sources":["../../src/utils/batches.ts"],"names":[],"mappings":"AAAA;;GAEG;AACH,wBAAgB,eAAe,CAAC,UAAU,EAAE,MAAM,GAAG,MAAM,EAAE,CAgE5D;AAED;;GAEG;AACH,wBAAgB,mBAAmB,CAAC,KAAK,EAAE,MAAM,GAAG,OAAO,CAG1D"}

package/dist/utils/batches.js

@@ -0,0 +1,61 @@
+/**
+ * Parse batch input to support ranges like "1-10" or "batch-1,batch-2"
+ */
+export function parseBatchInput(batchInput) {
+    // Check if it's a comma-separated list first
+    if (batchInput.includes(',')) {
+        const parts = batchInput
+            .split(',')
+            .map((b) => b.trim())
+            .filter((b) => b.length > 0);
+        const allBatches = [];
+        for (const part of parts) {
+            // Check if this part is a range
+            const rangeMatch = part.match(/^(\d+)-(\d+)$/);
+            if (rangeMatch) {
+                const start = parseInt(rangeMatch[1], 10);
+                const end = parseInt(rangeMatch[2], 10);
+                if (start > end) {
+                    throw new Error(`Invalid batch range: start (${start}) cannot be greater than end (${end})`);
+                }
+                if (end - start >= 100) {
+                    throw new Error(`Batch range too large: ${end - start + 1} batches. Maximum allowed: 100`);
+                }
+                for (let i = start; i <= end; i++) {
+                    allBatches.push(i.toString());
+                }
+            }
+            else {
+                // Single batch
+                allBatches.push(part);
+            }
+        }
+        return allBatches;
+    }
+    // Check if it's a single range (e.g., "1-10")
+    const rangeMatch = batchInput.match(/^(\d+)-(\d+)$/);
+    if (rangeMatch) {
+        const start = parseInt(rangeMatch[1], 10);
+        const end = parseInt(rangeMatch[2], 10);
+        if (start > end) {
+            throw new Error(`Invalid batch range: start (${start}) cannot be greater than end (${end})`);
+        }
+        if (end - start >= 100) {
+            throw new Error(`Batch range too large: ${end - start + 1} batches. Maximum allowed: 100`);
+        }
+        const batches = [];
+        for (let i = start; i <= end; i++) {
+            batches.push(i.toString());
+        }
+        return batches;
+    }
+    // Single batch
+    return [batchInput];
+}
+/**
+ * Validate batch format
+ */
+export function validateBatchFormat(batch) {
+    // Allow numeric batches (1, 2, 3) or named batches (batch-1, Batch_01, etc.)
+    return /^[\w\-_]+$/.test(batch);
+}

package/dist/utils/batches.test.d.ts

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=batches.test.d.ts.map

package/dist/utils/batches.test.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"batches.test.d.ts","sourceRoot":"","sources":["../../src/utils/batches.test.ts"],"names":[],"mappings":""}

package/dist/utils/batches.test.js

@@ -0,0 +1,119 @@
+import { describe, it, expect } from 'vitest';
+import { parseBatchInput, validateBatchFormat } from './batches.js';
+describe('Batch Utilities', () => {
+    describe('parseBatchInput', () => {
+        describe('single batches', () => {
+            it('should parse single numeric batch', () => {
+                expect(parseBatchInput('1')).toEqual(['1']);
+                expect(parseBatchInput('42')).toEqual(['42']);
+            });
+            it('should parse single named batch', () => {
+                expect(parseBatchInput('batch-1')).toEqual(['batch-1']);
+                expect(parseBatchInput('Batch_01')).toEqual(['Batch_01']);
+                expect(parseBatchInput('historical-content')).toEqual(['historical-content']);
+            });
+        });
+        describe('numeric ranges', () => {
+            it('should parse simple ranges', () => {
+                expect(parseBatchInput('1-3')).toEqual(['1', '2', '3']);
+                expect(parseBatchInput('5-10')).toEqual(['5', '6', '7', '8', '9', '10']);
+                expect(parseBatchInput('1-1')).toEqual(['1']);
+            });
+            it('should handle large ranges', () => {
+                const result = parseBatchInput('1-100');
+                expect(result).toHaveLength(100);
+                expect(result[0]).toBe('1');
+                expect(result[99]).toBe('100');
+            });
+            it('should reject invalid ranges', () => {
+                expect(() => parseBatchInput('10-5')).toThrow('Invalid batch range: start (10) cannot be greater than end (5)');
+                expect(() => parseBatchInput('5-5')).not.toThrow(); // Valid single-item range
+            });
+            it('should reject ranges that are too large', () => {
+                expect(() => parseBatchInput('1-102')).toThrow('Batch range too large: 102 batches. Maximum allowed: 100');
+                expect(() => parseBatchInput('1-101')).toThrow('Batch range too large: 101 batches. Maximum allowed: 100');
+                expect(() => parseBatchInput('1-100')).not.toThrow(); // Valid maximum range
+            });
+        });
+        describe('comma-separated lists', () => {
+            it('should parse simple comma-separated lists', () => {
+                expect(parseBatchInput('1,2,3')).toEqual(['1', '2', '3']);
+                expect(parseBatchInput('batch-1,batch-2,batch-3')).toEqual([
+                    'batch-1',
+                    'batch-2',
+                    'batch-3',
+                ]);
+            });
+            it('should handle whitespace in comma-separated lists', () => {
+                expect(parseBatchInput('1, 2, 3')).toEqual(['1', '2', '3']);
+                expect(parseBatchInput(' 1 , 2 , 3 ')).toEqual(['1', '2', '3']);
+            });
+            it('should filter out empty entries', () => {
+                expect(parseBatchInput('1,,2,3')).toEqual(['1', '2', '3']);
+                expect(parseBatchInput('1, ,2,3')).toEqual(['1', '2', '3']);
+            });
+        });
+        describe('mixed formats', () => {
+            it('should handle ranges within comma-separated lists', () => {
+                expect(parseBatchInput('1-3,5,7-9')).toEqual(['1', '2', '3', '5', '7', '8', '9']);
+                expect(parseBatchInput('1-5,10,15-17')).toEqual([
+                    '1',
+                    '2',
+                    '3',
+                    '4',
+                    '5',
+                    '10',
+                    '15',
+                    '16',
+                    '17',
+                ]);
+            });
+            it('should handle complex mixed formats', () => {
+                expect(parseBatchInput('1-3,batch-1,5-7,historical')).toEqual([
+                    '1',
+                    '2',
+                    '3',
+                    'batch-1',
+                    '5',
+                    '6',
+                    '7',
+                    'historical',
+                ]);
+            });
+        });
+        describe('edge cases', () => {
+            it('should handle empty string', () => {
+                expect(parseBatchInput('')).toEqual(['']);
+            });
+            it('should handle single comma', () => {
+                expect(parseBatchInput(',')).toEqual([]);
+            });
+            it('should handle multiple commas', () => {
+                expect(parseBatchInput(',,')).toEqual([]);
+                expect(parseBatchInput('1,,2')).toEqual(['1', '2']);
+            });
+        });
+    });
+    describe('validateBatchFormat', () => {
+        it('should accept valid batch names', () => {
+            expect(validateBatchFormat('1')).toBe(true);
+            expect(validateBatchFormat('42')).toBe(true);
+            expect(validateBatchFormat('batch-1')).toBe(true);
+            expect(validateBatchFormat('Batch_01')).toBe(true);
+            expect(validateBatchFormat('historical-content')).toBe(true);
+            expect(validateBatchFormat('content_2023')).toBe(true);
+        });
+        it('should reject invalid batch names', () => {
+            expect(validateBatchFormat('')).toBe(false);
+            expect(validateBatchFormat('batch 1')).toBe(false); // space not allowed
+            expect(validateBatchFormat('batch.1')).toBe(false); // dot not allowed
+            expect(validateBatchFormat('batch/1')).toBe(false); // slash not allowed
+            expect(validateBatchFormat('batch@1')).toBe(false); // @ not allowed
+        });
+        it('should handle special characters correctly', () => {
+            expect(validateBatchFormat('batch-1')).toBe(true); // hyphen allowed
+            expect(validateBatchFormat('batch_1')).toBe(true); // underscore allowed
+            expect(validateBatchFormat('Batch01')).toBe(true); // alphanumeric allowed
+        });
+    });
+});

package/dist/utils/default-server.d.ts

@@ -0,0 +1,3 @@
+export declare function getCliName(): 'openrxiv' | 'biorxiv' | 'medrxiv';
+export declare function getDefaultServer(): 'biorxiv' | 'medrxiv';
+//# sourceMappingURL=default-server.d.ts.map

package/dist/utils/default-server.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"default-server.d.ts","sourceRoot":"","sources":["../../src/utils/default-server.ts"],"names":[],"mappings":"AAEA,wBAAgB,UAAU,IAAI,UAAU,GAAG,SAAS,GAAG,SAAS,CAY/D;AAED,wBAAgB,gBAAgB,IAAI,SAAS,GAAG,SAAS,CAOxD"}

package/dist/utils/default-server.js

@@ -0,0 +1,20 @@
+import path from 'node:path';
+export function getCliName() {
+    // process.argv[1] contains the script path, which includes the alias
+    const scriptPath = process.argv[1];
+    const commandName = path.basename(scriptPath);
+    if (commandName.toLowerCase().includes('biorxiv')) {
+        return 'biorxiv';
+    }
+    if (commandName.toLowerCase().includes('medrxiv')) {
+        return 'medrxiv';
+    }
+    return 'openrxiv';
+}
+export function getDefaultServer() {
+    const cliName = getCliName();
+    if (cliName.toLowerCase().includes('medrxiv')) {
+        return 'medrxiv';
+    }
+    return 'biorxiv';
+}

package/dist/utils/index.d.ts

@@ -0,0 +1,5 @@
+export * from './months.js';
+export * from './meca-processor.js';
+export * from './batches.js';
+export * from './default-server.js';
+//# sourceMappingURL=index.d.ts.map

package/dist/utils/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/utils/index.ts"],"names":[],"mappings":"AACA,cAAc,aAAa,CAAC;AAC5B,cAAc,qBAAqB,CAAC;AACpC,cAAc,cAAc,CAAC;AAC7B,cAAc,qBAAqB,CAAC"}

package/dist/utils/index.js

@@ -0,0 +1,5 @@
+// Export all utility functions
+export * from './months.js';
+export * from './meca-processor.js';
+export * from './batches.js';
+export * from './default-server.js';

package/dist/utils/meca-processor.d.ts

@@ -0,0 +1,28 @@
+export interface ProcessMecaOptions {
+    batch: string;
+    server: 'biorxiv' | 'medrxiv';
+    apiUrl: string;
+    output?: string;
+    s3Key: string;
+    apiKey?: string;
+    selective?: boolean;
+}
+export interface ProcessMecaResult {
+    success: boolean;
+    paper?: any;
+    error?: string;
+}
+/**
+ * Process a MECA file and extract metadata
+ * @param mecaPath Path to the MECA file (local file path)
+ * @param options Processing options
+ * @returns ProcessMecaResult with success status and extracted paper data
+ */
+export declare function processMecaFile(mecaPath: string, options: ProcessMecaOptions): Promise<ProcessMecaResult>;
+/**
+ * Preprocess XML content to fix common HTML entities that cause parsing errors
+ * @param xmlContent Raw XML content
+ * @returns Preprocessed XML content with entities replaced
+ */
+export declare function preprocessXMLContent(xmlContent: string): string;
+//# sourceMappingURL=meca-processor.d.ts.map

package/dist/utils/meca-processor.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"meca-processor.d.ts","sourceRoot":"","sources":["../../src/utils/meca-processor.ts"],"names":[],"mappings":"AAqCA,MAAM,WAAW,kBAAkB;IACjC,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,SAAS,GAAG,SAAS,CAAC;IAC9B,MAAM,EAAE,MAAM,CAAC;IACf,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,SAAS,CAAC,EAAE,OAAO,CAAC;CACrB;AAED,MAAM,WAAW,iBAAiB;IAChC,OAAO,EAAE,OAAO,CAAC;IACjB,KAAK,CAAC,EAAE,GAAG,CAAC;IACZ,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAED;;;;;GAKG;AACH,wBAAsB,eAAe,CACnC,QAAQ,EAAE,MAAM,EAChB,OAAO,EAAE,kBAAkB,GAC1B,OAAO,CAAC,iBAAiB,CAAC,CAkE5B;AAwQD;;;;GAIG;AACH,wBAAgB,oBAAoB,CAAC,UAAU,EAAE,MAAM,GAAG,MAAM,CAwD/D"}