Download and extract the parser source code into some folder, for example to folder “D:/qparserSrc”.
Follow the below steps to setup the python environment.
-
Install the 64-Bit version of Miniconda from (https://repo.anaconda.com/miniconda/).
-
Open a new window of anaconda prompt. Create a new conda environment with a name “D:/condaEnv/qparserSOA”. This creates the conda environment inside the folder “D:/condaEnv/qparserSOA”.
conda create -p D:/condaEnv/qparserSOA python=3.9.7
- Activate the new environment:
conda activate D:/condaEnv/qparserSOA
- Move to the folder qparserSOA. Packages from github will be installed here.
d:
cd D:\condaEnv\qparserSOA
- Install spacy package from conda-forge
conda install -c conda-forge spacy
- Install spacy trained pipeline. If the installation throws error, then try executing the command again.
python -m spacy download en_core_web_sm
- Install other packages from conda-forge:
conda install -c conda-forge antlr4-python3-runtime=4.9.3 word2number pyzmq
- Optionally, it may be necessary to install the checklist package:
pip install checklist
- In the anaconda prompt, make sure the “D:/condaEnv/qparserSOA” environment is activated.
- Move to the folder “D:/qparserSrc/test”. Packages from github will be installed here.
cd D:\qparserSrc\test
- Execute the "Test.py" script:
python Test.py
- The output of the script should look like below:
============================== TESTING retrieval dataset
============================== TESTING GeoAnQu dataset
============================== TESTING MS student dataset
Process finished with exit code 0
- Make sure “D:/condaEnv/qparserSOA” is activated in the anaconda prompt.
- In the anaconda prompt, navigate to the folder “D:/qparserSrc”. This folder should contain the batch file “runAsyncWorker.bat”.
- Run the batch file by typing
runAsyncWorker.bat
- If the server starts correctly then you should see a message like below:
error loading _jsonnet (this is expected on Windows), treating config.json as plain json
INFO: Bound broker frontend to 'tcp://127.0.0.1:5570' in method 'QparserBroker.run'
INFO: Bound broker backend to an inter-process port 'inproc://backend' in method 'QparserBroker.run'
INFO: Started worker '0' on a inter-process socket 'inproc://backend' in method 'QparserWorker.run'
INFO: Started worker '1' on a inter-process socket 'inproc://backend' in method 'QparserWorker.run'
INFO: Started worker '2' on a inter-process socket 'inproc://backend' in method 'QparserWorker.run'
INFO: Started the poller in the broker in method 'QparserBroker.run'
To stop the server press the combo “Ctrl+C” in the anaconda prompt. Enter ‘Y’ when asked “Terminate batch job (Y/N)?”.
Two parameters can be changed in the batch file “runAsyncWorker.bat”. “FRONT_PORT” sets to which port server should be bound to. “INST_COUNT” sets the number of concurrent worker threads. In other words, it is the number of requests the server can handle simultaneously without requests blocking each other. For example, if “INST_COUNT” is set to 1 then only one request is processed at the time, and all other incoming requests are queued until the current request is handled. “INST_COUNT” can be any integer number above 0.
- Make sure the parser server is up and running.
- Open a new Anaconda prompt window.
- In the newly opened prompt window, activate the environment “D:/condaEnv/qparserSOA”.
- Navigate to the folder “D:/qparserSrc”.
- Execute the "asyncClient.py" script:
- The expected output should look like below:
Starting the client
Setting the client poller
Sending a request to the remove service
Waiting for a reply ...
Client received a reply: {"question": "What is the shortest path through my workplace, a gym and a supermarket
from my home in Amsterdam", "placename": ["Amsterdam"], "replaceQ": "What is shortest path through workplace ,
gym and supermarket from home extent", "network": ["shortest path"], "object": ["supermarket", "workplace",
"home", "gym"], "ner_Question": "what is network0 through object1 , object3 and object0 from object2 extent",
"parseTreeStr": "(start what is (measure (networkC network 0) through (destination (objectC object 1) (objectC
object 3) and (objectC object 0)) from (origin (objectC object 2))) (extent extent))", "cctrans": {"types":
[{"type": "object", "id": "0", "keyword": "home", "cct": "R(Obj,_)"}, {"type": "object", "id": "1", "keyword":
"supermarket", "cct": "R(Obj,_)"}, {"type": "object", "id": "2", "keyword": "gym", "cct": "R(Obj,_)"}, {"type":
"object", "id": "3", "keyword": "workplace", "cct": "R(Obj,_)"}, {"type": "network", "id": "4", "keyword":
"shortest path", "cct": "R(Obj*Obj,Reg)"}, {"type": "object", "id": "5", "keyword": "Amsterdam", "cct":
"R(Obj,_)"}], "extent": ["5"], "transformations": [{"before": ["1", "2", "3", "0"], "after": ["4"]}]}, "valid":
"T", "query": {"afterId": "4", "after": "R(Obj*Obj,Reg)", "before": ["R(Obj,_)", "R(Obj,_)", "R(Obj,_)",
"R(Obj,_)"]}, "queryEx": {"after": {"id": "4", "cct": "R(Obj*Obj,Reg)"}, "before": [{"after": {"id": "1", "cct":
"R(Obj,_)"}}, {"after": {"id": "2", "cct": "R(Obj,_)"}}, {"after": {"id": "3", "cct": "R(Obj,_)"}}, {"after":
{"id": "0", "cct": "R(Obj,_)"}}]}}
Process finished with exit code 0
from QuestionParser import *
from TypesToQueryConverter import *
qBlock = '{"question": "What is the shortest path through my workplace,' + \
' a gym and a supermarket from my home in Amsterdam","placename":' + \
'["Amsterdam"],"replaceQ": "What is the shortest path through my' + \
' workplace, a gym and a supermarket from my home extent"}'
# identify types and transformations steps
parser = QuestionParser()
qParsed = parser.parseQuestionBlock(qBlock)
# annotate types with cct expressions and generate a query
cctAnnotator = TQConverter()
cctAnnotator.cctToExpandedQuery(qParsed, True, True)Check the Grammar folder for more details of the functional grammar in GeoAnQu.g4.
Check the Dictionary folder for the concept dictionary.