updated readme

README.md

@@ -16,6 +16,67 @@ The user provides a collection of documents for topic analysis and the service w
 consists of a collection of words that represent a given topic.
 
 
+All services have been tested to work with python3.6.
+
+
+
+## User Guide
+
+Please look at the [user guide](docs/USERGUIDE.md) for a detailed spec of the services and how to use the services.
+
+
+
+## Running the service locally
+
+### Install preprequisites
+
+```
+pip install -r requirements.txt
+```
+
+
+### Setup
+
+Run the following commands to generate gRPC classes for Python
+
+```
+python3.6 -m grpc_tools.protoc -I. --python_out=.  --grpc_python_out=. service_spec/topic_analysis.proto
+```
+
+
+
+### Running unit tests
+
+
+```
+python3.6 test_topic_analysis_grpc.py
+```
+
+### Usage
+
+To start the gRPC server locally
+
+```
+python3.6 topic_analysis_grpc.py
+
+```
+
+Topic analysis would most likely involve running a dimensionality reduction alogrithm which would take a considerable lenght of time to complete. For this reason, running the service above would return a handle which you will need to use to query a restapi endpoint at a later time. See the user guide for details. You can start that endpoint with the command
+
+```
+python3.6 analysis_results.py
+```
+
+You can also use a suitable application server for python flask. A sample config file using gunicorn is [config.py](Docker/gunicorn/config.py).
+
+Then, you can execute the below command to serve analysis_results.py, by executing gnunicorn in a folder containing config.py, while the configuration file confi.py has a path pointing to analysis_results.py.
+
+```
+gunicorn -c config.py analysis_results:app
+```
+
+
+
 ## Resources
 
 LSA:
@@ -31,4 +92,15 @@ LDA:
  * [Wikipedia entry](https://en.wikipedia.org/wiki/Latent_Dirichlet_allocation)
 
 LDA2vec:
- * Research paper: [Mixing Dirichlet Topic Models and Word Embeddings to Make lda2vec](https://arxiv.org/abs/1605.02019)
+ * Research paper: [Mixing Dirichlet Topic Models and Word Embeddings to Make lda2vec](https://arxiv.org/abs/1605.02019)
+
+
+## Contributors
+
+### Authors
+
+* Eyob Yirdaw
+
+### Maintainers
+
+* Eyob Yirdaw

docs/USERGUIDE.md

@@ -0,0 +1,22 @@
+[![SingnetLogo](../docs/assets/singnet-logo.jpg?raw=true 'SingularityNET')](https://singularitynet.io/)
+
+
+
+Below is available list of topic analysis methods that have been implemented so far.
+
+
+## Probabilistic Latent Semantic Analysis (PLSA)
+
+Existing parameters are
+
+- **docs**: a collection of documents, either supplied as an array of strings or a collection of .txt files. At least two documents should be given. Note that a single document can consist of a single sentence. In this way
+you can extract topics from a single text file by first tokenizing it at the sentence level and treating each sentence as a single "document".
+- **num_topics**: The number of topics to extract. The minimum value is 2.
+- **topic_divider**: This value is an integer value. If it has a value of zero, then num_topics would be used. If it has a positive value, then the number of topics would be
+the `(number of documents)/topic_divider`
+- **maxiter**: This value gives the maximum EM (expectation maximum algorithm) iteration that is allowed. Although a suitable value depends on the supplied documents and the value of beta, a good value can be 22.
+- **beta**: This is a floating point number with range of (0,1]. It is used as a tempering parameter in the EM algorithm. A recommended values are values closer to 1 including 1 itself. Choosing the value of 1 means no tempering is used.
+
+
+Here are sample json files to try with this algorithm, either from the snet client or the dApp: [longer sample](../docs/tests/topic_analysis.json), [shorter sample](../docs/tests/topic_analysis_2.json) .
+