README Edits, Small utils.py Edit

.gitignore

@@ -105,3 +105,7 @@ venv.bak/
 
 .DS_Store
 .idea/
+
+#sean_fork
+sean_notes.txt
+results/*

CO_ATTN/utils.py

@@ -75,7 +75,7 @@ def load_word_embedding_dict(embedding, embedding_path, word_alphabet, logger, e
         logger.info("Loading GloVe ...")
         embedd_dim = -1
         embedd_dict = dict()
-        with open(embedding_path, 'r') as file:
+        with open(embedding_path, 'r', encoding='utf-8') as file:
             for line in file:
                 line = line.strip()
                 if len(line) == 0:

README.md

@@ -1,22 +1,54 @@
-# co-attention
+# Co-Attention
 Code for BEA 13 paper "Co-Attention Based Neural Network for Source-Dependent Essay Scoring"
 
 ## Dependencies
 
-python 2 for data/preprocess_asap.py (will be upgraded to python 3)
+Python 2 for data/preprocess_asap.py (will be upgraded to Python 3).
 
-python 3 for the rest
+* I recommend that on installation, you *do not add to your PATH variable*. This way, it doesn't interfere with your current Python workflow.
+* Then, when you need to run the preprocessing script, you'll run it something like:
+	* *c:/Python27/python.exe preprocess_asap.py*	
+
+Python 3 for the rest
 
 * tensorflow 2.0.0 beta
 * gensim
+	* gensim may have more dependencies, such as VS tools
 * nltk
 * sklearn
 
-run python2 data/preprocess_asap.py for data splitting.
-Download Glove pretrained embedding from https://nlp.stanford.edu/projects/glove
-Extract glove.6B.50d.txt to the glove folder
-run python3 attn_network.py [options] for training and evaluation
+## Running on Linux, MacOS
+
+1. Run *python2 data/preprocess_asap.py* for data splitting.
+2. Download GloVe pretrained embedding from: *https://nlp.stanford.edu/projects/glove*
+3. Extract *glove.6B.50d.txt* to the glove folder
+4. Run *python3 attn_network.py [options]* for training and evaluation
+
+## Running on Windows
+
+To run on Windows, do all of the commands for Linux/MacOS. Then you'll need to remove two "\n" symbols from the preprocessing script.
+
+1. open *data/preprocess_asap.py* in your preferred text editor
+2. on lines 28 and 31 in the preprocessing script, you'll find: *f_write.write("\r\n")*
+3. remove the *\n* from both lines
+
+## Results
+
+After preprocessing the data, the program will start the training process. At the end of each epoch, the logger will
+output the development and testing set scores. The highest will be kept and outputted after all epochs are complete. You can toggle the which
+task (the default is ASAP3), the number of epochs (the default is 50), and more by looking at the arguments in lines 21-51 of *attn_network.py*.
+
+Additionally, if you want to look at specific essays with their predicted and actual scores:
+
+1. go to the *checkpoints* folder
+2. after training, there should be a text file with one number per line. the line number corresponds to the essay number in the test data.
+3. in the various fold directories, open the test.tsv file and compare with the predicted scored in from step 2.
+
+## Coming Soon
 
+1. making specific essays, as well as their predicted scores and real scores, more accessible. 
+	- likely a Python script
+2. updating *the preprocessing_asap.py* to be compatible with Python 3
 
 ## Cite
 If you use the code, please cite the following paper:

requirements.txt

@@ -0,0 +1,46 @@
+absl-py==0.11.0
+astunparse==1.6.3
+cachetools==4.2.0
+certifi==2020.12.5
+chardet==4.0.0
+click==7.1.2
+Cython==0.29.14
+flatbuffers==1.12
+gast==0.3.3
+gensim==3.8.3
+google-auth==1.24.0
+google-auth-oauthlib==0.4.2
+google-pasta==0.2.0
+grpcio==1.32.0
+h5py==2.10.0
+idna==2.10
+joblib==1.0.0
+Keras-Preprocessing==1.1.2
+Markdown==3.3.3
+nltk==3.5
+numpy==1.19.5
+oauthlib==3.1.0
+opt-einsum==3.3.0
+protobuf==3.14.0
+pyasn1==0.4.8
+pyasn1-modules==0.2.8
+regex==2020.11.13
+requests==2.25.1
+requests-oauthlib==1.3.0
+rsa==4.7
+scikit-learn==0.24.1
+scipy==1.6.0
+six==1.15.0
+sklearn==0.0
+smart-open==4.1.2
+tensorboard==2.4.1
+tensorboard-plugin-wit==1.8.0
+tensorflow==2.4.0
+tensorflow-estimator==2.4.0
+termcolor==1.1.0
+threadpoolctl==2.1.0
+tqdm==4.56.0
+typing-extensions==3.7.4.3
+urllib3==1.26.2
+Werkzeug==1.0.1
+wrapt==1.12.1