Add various tools for enhanced functionality

Author: matyo91

src/agent/src/Toolbox/Tool/Airtable.php

@@ -0,0 +1,257 @@
+<?php
+
+/*
+ * This file is part of the Symfony package.
+ *
+ * (c) Fabien Potencier <[email protected]>
+ *
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ */
+
+namespace Symfony\AI\Agent\Toolbox\Tool;
+
+use Symfony\AI\Agent\Toolbox\Attribute\AsTool;
+use Symfony\AI\Platform\Contract\JsonSchema\Attribute\With;
+use Symfony\Contracts\HttpClient\HttpClientInterface;
+
+/**
+ * @author Mathieu Ledru <[email protected]>
+ */
+#[AsTool('arxiv_search', 'Tool that searches the ArXiv API for scientific papers')]
+final readonly class ArXiv
+{
+    /**
+     * @param array<string, mixed> $options Additional search options
+     */
+    public function __construct(
+        private HttpClientInterface $httpClient,
+        private array $options = [],
+    ) {
+    }
+
+    /**
+     * @param string $query      search query to look up on ArXiv
+     * @param int    $maxResults The maximum number of results to return
+     * @param int    $start      The starting position for results (for pagination)
+     *
+     * @return array<int, array{
+     *     id: string,
+     *     title: string,
+     *     authors: array<int, string>,
+     *     summary: string,
+     *     published: string,
+     *     updated: string,
+     *     categories: array<int, string>,
+     *     link: string,
+     * }>
+     */
+    public function __invoke(
+        #[With(maximum: 500)]
+        string $query,
+        int $maxResults = 10,
+        #[With(minimum: 0)]
+        int $start = 0,
+    ): array {
+        try {
+            // Check if query is an ArXiv identifier
+            if ($this->isArxivIdentifier($query)) {
+                return $this->searchById($query);
+            }
+
+            return $this->searchByQuery($query, $maxResults, $start);
+        } catch (\Exception $e) {
+            return [
+                [
+                    'id' => 'error',
+                    'title' => 'Search Error',
+                    'authors' => [],
+                    'summary' => 'Unable to perform ArXiv search: '.$e->getMessage(),
+                    'published' => '',
+                    'updated' => '',
+                    'categories' => [],
+                    'link' => '',
+                ],
+            ];
+        }
+    }
+
+    /**
+     * Check if a query is an ArXiv identifier.
+     */
+    private function isArxivIdentifier(string $query): bool
+    {
+        // ArXiv identifier patterns
+        $patterns = [
+            '/^\d{4}\.\d{4,5}(v\d+)?$/',  // YYMM.NNNN or YYMM.NNNNvN
+            '/^\d{7}\.\d+$/',             // YYMMNNN.N
+            '/^[a-z-]+\/\d{7}$/',         // category/YYMMNNN
+        ];
+
+        foreach ($patterns as $pattern) {
+            if (preg_match($pattern, $query)) {
+                return true;
+            }
+        }
+
+        return false;
+    }
+
+    /**
+     * Search ArXiv by specific paper ID.
+     *
+     * @return array<int, array{
+     *     id: string,
+     *     title: string,
+     *     authors: array<int, string>,
+     *     summary: string,
+     *     published: string,
+     *     updated: string,
+     *     categories: array<int, string>,
+     *     link: string,
+     * }>
+     */
+    private function searchById(string $id): array
+    {
+        $response = $this->httpClient->request('GET', 'http://export.arxiv.org/api/query', [
+            'query' => [
+                'id_list' => $id,
+                'max_results' => 1,
+            ],
+        ]);
+
+        $xml = $response->getContent();
+
+        return $this->parseArxivXml($xml);
+    }
+
+    /**
+     * Search ArXiv by query string.
+     *
+     * @return array<int, array{
+     *     id: string,
+     *     title: string,
+     *     authors: array<int, string>,
+     *     summary: string,
+     *     published: string,
+     *     updated: string,
+     *     categories: array<int, string>,
+     *     link: string,
+     * }>
+     */
+    private function searchByQuery(string $query, int $maxResults, int $start): array
+    {
+        $response = $this->httpClient->request('GET', 'http://export.arxiv.org/api/query', [
+            'query' => array_merge($this->options, [
+                'search_query' => $query,
+                'start' => $start,
+                'max_results' => $maxResults,
+                'sortBy' => 'relevance',
+                'sortOrder' => 'descending',
+            ]),
+        ]);
+
+        $xml = $response->getContent();
+
+        return $this->parseArxivXml($xml);
+    }
+
+    /**
+     * Parse ArXiv XML response.
+     *
+     * @return array<int, array{
+     *     id: string,
+     *     title: string,
+     *     authors: array<int, string>,
+     *     summary: string,
+     *     published: string,
+     *     updated: string,
+     *     categories: array<int, string>,
+     *     link: string,
+     * }>
+     */
+    private function parseArxivXml(string $xml): array
+    {
+        try {
+            $dom = new \DOMDocument();
+            $dom->loadXML($xml);
+
+            $entries = $dom->getElementsByTagName('entry');
+            $results = [];
+
+            foreach ($entries as $entry) {
+                $id = $this->getElementText($entry, 'id');
+                $title = $this->getElementText($entry, 'title');
+                $summary = $this->getElementText($entry, 'summary');
+                $published = $this->getElementText($entry, 'published');
+                $updated = $this->getElementText($entry, 'updated');
+
+                // Extract authors
+                $authors = [];
+                $authorNodes = $entry->getElementsByTagName('author');
+                foreach ($authorNodes as $authorNode) {
+                    $name = $this->getElementText($authorNode, 'name');
+                    if ($name) {
+                        $authors[] = $name;
+                    }
+                }
+
+                // Extract categories
+                $categories = [];
+                $categoryNodes = $entry->getElementsByTagName('category');
+                foreach ($categoryNodes as $categoryNode) {
+                    $term = $categoryNode->getAttribute('term');
+                    if ($term) {
+                        $categories[] = $term;
+                    }
+                }
+
+                // Extract link
+                $link = '';
+                $linkNodes = $entry->getElementsByTagName('link');
+                foreach ($linkNodes as $linkNode) {
+                    if ('pdf' === $linkNode->getAttribute('title')) {
+                        $link = $linkNode->getAttribute('href');
+                        break;
+                    }
+                }
+
+                $results[] = [
+                    'id' => $id,
+                    'title' => $title,
+                    'authors' => $authors,
+                    'summary' => trim($summary),
+                    'published' => $published,
+                    'updated' => $updated,
+                    'categories' => $categories,
+                    'link' => $link,
+                ];
+            }
+
+            return $results;
+        } catch (\Exception $e) {
+            return [
+                [
+                    'id' => 'parse_error',
+                    'title' => 'Parse Error',
+                    'authors' => [],
+                    'summary' => 'Unable to parse ArXiv response: '.$e->getMessage(),
+                    'published' => '',
+                    'updated' => '',
+                    'categories' => [],
+                    'link' => '',
+                ],
+            ];
+        }
+    }
+
+    private function getElementText(\DOMElement $parent, string $tagName): string
+    {
+        $elements = $parent->getElementsByTagName($tagName);
+        if ($elements->length > 0) {
+            return $elements->item(0)->textContent ?? '';
+        }
+
+        return '';
+    }
+}