This project do relation extraction work by Convolutional Neural Network
The tool we used is Tensorflow

1. pre-processing python file of ACE data, SemEval data and Chinese data
   
2. python file for relation extraction by CNN architecture
   

preceprocess_ace.py -- main function to extraction structure data from ace and save it as embedding form
c_xml_parse.py -- class for parse xml
c_ace_process.py -- classes for ACE Data process, includes definition of data structure, and parse function
c_relation.py -- class for turning relation mention into embedding
c_open_tool.py -- class for other tool, includes load position matrix and word matrix and so on.

data2vector.py -- load the trained word2vector embed
load_data.py -- word segmentation and PF value counting, to change the raw data to [word_id, pf_id, relation_id] form

seg_tool:
cws.model is the segmentation model provided by LTP
vocab.txt is the dictionary which can be used in segmentation
word2vec:
vectors.txt is the trained chinese word embedding by word2vector tool

the relation mention we extracted from raw data is the form as follow:
Ivan in Wisconsin PHYS Located
entity are tagged followed by relation tag