From 464b8b04ea0d51280849173d5eda92d4d4db8612 Mon Sep 17 00:00:00 2001 From: roryhaung Date: Wed, 16 Oct 2024 20:05:36 +0800 Subject: [PATCH] test: Add smart_scraper_multi_parse_merge_first_graph test --- ...aper_multi_parse_merge_first_graph_test.py | 59 +++++++++++++++++++ 1 file changed, 59 insertions(+) create mode 100644 tests/graphs/smart_scraper_multi_parse_merge_first_graph_test.py diff --git a/tests/graphs/smart_scraper_multi_parse_merge_first_graph_test.py b/tests/graphs/smart_scraper_multi_parse_merge_first_graph_test.py new file mode 100644 index 00000000..506ce5da --- /dev/null +++ b/tests/graphs/smart_scraper_multi_parse_merge_first_graph_test.py @@ -0,0 +1,59 @@ +""" +Module for testing the smart scraper class +""" + +import os +import pytest +import pandas as pd +from dotenv import load_dotenv +from scrapegraphai.graphs import SmartScraperMultiParseConcatFirstGraph +from scrapegraphai.utils import prettify_exec_info + +load_dotenv() + +@pytest.fixture +def graph_config(): + """Configuration of the graph""" + openai_key = os.getenv("OPENAI_APIKEY") + + return { + "llm": { + "api_key": openai_key, + "model": "openai/gpt-3.5-turbo", + }, + "verbose": True, + "headless": False, + } + +def test_scraping_pipeline(graph_config): + """Start of the scraping pipeline""" + smart_scraper_multi_parse_concat_first_graph = SmartScraperMultiParseConcatFirstGraph( + prompt="Who is Marco Perini?", + source= [ + "https://perinim.github.io/", + "https://perinim.github.io/cv/" + ], + config=graph_config, + ) + + result = smart_scraper_multi_parse_concat_first_graph.run() + + assert result is not None + assert isinstance(result, dict) + +def test_get_execution_info(graph_config): + """Get the execution info""" + smart_scraper_multi_parse_concat_first_graph = SmartScraperMultiParseConcatFirstGraph( + prompt="Who is Marco Perini?", + source= [ + "https://perinim.github.io/", + "https://perinim.github.io/cv/" + ], + config=graph_config, + ) + + smart_scraper_multi_parse_concat_first_graph.run() + + graph_exec_info = smart_scraper_multi_parse_concat_first_graph.get_execution_info() + + assert graph_exec_info is not None