Commit 796d8a43 authored by Nandeesh Patel Gowdru Prabushanker's avatar Nandeesh Patel Gowdru Prabushanker
Browse files

Updating the documentation

parent 90d17385
......@@ -75,3 +75,11 @@ The deployed webservice does not reflect the optimal parametrization of AGDISTIS
mvn clean package tomcat:run
```
For more information, go to our <a href="https://github.com/AKSW/AGDISTIS/wiki/3-Running-the-webservice">Wiki</a>.
### Usage for "Entity Linking using Neural Networks" thesis
1. Download the gerbil datasets from https://github.com/AKSW/gerbil/releases/download/v1.2.5/gerbil_data.zip
2. Extract the contents and change the parentPath value in GetCandidates class Main function
3. Run ```mvn clean & mvn install & mvn exec:java -Dexec.mainClass="org.aksw.agdistis.webapp.GetCandidates``` to generate the training files needed for the model
4. Run ```mvn exec:java -Dexec.mainClass="org.aksw.agdistis.webapp.RunApp``` to start the server
\ No newline at end of file
## DBpedia entity embeddings
This repo has APIs the can be used to access embeddings for all the DBpedia entities. The embeddings are indexed on Elasticsearch server.
<br><br>
This repo contains source code for the thesis "Entity Linking using Neural Networks".
### List of APIs
Using the repo to link entities in a text
1. Use AGDISTIS to create training files as explained in the documentation
2. Install all dependencies from requirements.txt
3. Run ```python model.py``` to train and save entity linking models
4. Run the Flask app using ```python run.py```
### List of APIs available
#### 1. Get entity embeddings
This API takes a list of entities as input and returns the embeddings of the given entities in response. It returns embeddings of first 10 unique entities and ignores the rest.
......@@ -57,4 +63,6 @@ This API accepts a NIF document with a text and named entities existing in the t
Request Body: NIF content
```
These APIs can be accessed from http://nel.cs.upb.de:5000/ on UPB network for now.
\ No newline at end of file
Currently model.py file trains and saves all models used for evaluation. You can modify the code as explained in the comments of the file to use a single model.
The APIs can be accessed from http://localhost:5000/
\ No newline at end of file
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment