This is our final project for 2020 Fall EE208.🙂
In this project, we have crawled over ten thousands pages and six thousands photos and faces.
Then we created index and used Flask to build a web framework. And we used bootstrap framework to beautify the front page.
We have achieved searching by keywords,images or faces. And the results can be sorted by relevance or date. And the result photo will be painted on the page by an ai-drawer.
Besides, we use clustering methods k-means to analyze the news.
In static, it's our static resource for the front web, which includes the image and faces database and css js files.
Before using it, you should first add the static path to the app.py.
In templates, it's our html templates.
In genetic-drawing, it's a ai-drawing achieved by genetic-algorithm. It can imitate the way humans draw and produce a gif image of the search result. However, we didn't write it alone. It's an open source project from GitHub. We improve and adjust it.
sports_html and sohu_html are our original data.
sports_index is the index created by lucene.
app.py is used to create our web.
extract_face.py is used to extract all faces in a photo and create the faces database.
you would need the following python 3 libraries:
- opencv 3.4.1
- numpy 1.16.2
- matplotlib 3.0.3
- jieba 0.42.1
- Flask 1.1.2
- dlib 19.21.1
- lucene The project is runned in docker enviroment.
open the project in docker enviroment, and enter
python app.py