Changes for ironspeed.

.gitignore

@@ -1,6 +1,7 @@
 **/adverb-venv/
 **/models/**
 **/.vscode
+**/src/**
 
 
 

README.md

@@ -10,52 +10,56 @@ Download and install Anaconda from the official website https://docs.anaconda.co
 
 ### Installation
 
-Clone the project
+#### Clone the project
 ```
 git clone https://github.com/giganticode/adverb-backend.git
 ```
 
-#### Windows
-Run the 'install_and_run.bat'-Script - it will create a virtual conda environment called 'adverb', install all dependencies and start the webservice.
+#### Create conda environment 'adverb'
+For GPU-use:
 ```
-cd adverb-backend
-.\install_and_run.bat
+conda env create -f conda_env.yml
 ```
-
-#### Unix
-Run the 'install_and_run.sh'-Script - it will create a virtual conda environment called 'adverb', install all dependencies and start the webservice.
+For CPU-use:
 ```
-cd adverb-backend
-sh install_and_run.sh
-```
-
-### Manual installation
-
-#### Create conda environment 'adverb'
-```
-conda env create -f environment.yml
+conda env create -f conda_env_cpu.yml
 ```
 
 #### Activate the conda environment
+For GPU-use:
 ```
 conda activate adverb
 ```
+For CPU-use:
+```
+conda activate adverb-cpu
+```
 
-### Start the webservice
+### Manually start the webservice
 ```
 python webservice.py 
 ```
 
 Optional parameters:
 
---port=8080 => Set the port of the webservice (default 8080)
+--port=8090 => Set the port of the webservice (default 8090)
 
---host=127.0.0.1 => Set the host of the webservice (default 127.0.0.1 )
+--host=127.0.0.1 => Set the host of the webservice (default 0.0.0.0)
 
 --debug => Enable the debug mode (default false)
 
 
-#### Start the webservice with the provided script
-run the run.bat-script (Windows) or the run.sh-script (UNIX)
+### Start the webservice with the provided script
+Running the following script will start the webservice.
+Windows
+```
+run.bat
+``` 
+Unix
+``` 
+run.sh
+``` 
+
+## CPU execution
 
-Running this script will start the webservice.
+Install the conda environment from the `conda_env_cpu.yml` file. In addition, note that if you are testing CPU execution on a machine that includes GPUs you might need to specify CUDA_VISIBLE_DEVICES="" as part of your command.

conda_env.yml

@@ -0,0 +1,25 @@
+name: adverb
+channels:
+  - pytorch
+  - nvidia
+  - conda-forge
+  - defaults
+dependencies:
+  - python=3.7
+  - cudatoolkit=11.1
+  - faiss-gpu=1.7.0
+  - flask
+  - flask-cors
+  - pytorch=1.9.0
+  - torchaudio
+  - torchvision
+  - pip=21.0
+  - pip:
+    - "--editable=git+https://github.com/stanford-futuredata/ColBERT.git@fast_search#egg=colbert"
+    - bitarray
+    - gitpython
+    - scipy
+    - spacy
+    - tqdm
+    - transformers==4.10.0
+    - ujson

conda_env_cpu.yml

@@ -0,0 +1,20 @@
+name: adverb-cpu
+channels:
+  - pytorch
+  - conda-forge
+  - defaults
+dependencies:
+  - python=3.7
+  - flask
+  - flask-cors
+  - pytorch-cpu
+  - pip=21.0
+  - pip:
+    - "--editable=git+https://github.com/stanford-futuredata/ColBERT.git@fast_search#egg=colbert"
+    - bitarray
+    - gitpython
+    - scipy
+    - spacy
+    - tqdm
+    - transformers==4.10.0
+    - ujson

controllers/code_search_colbert_controller.py

@@ -1,3 +1,4 @@
+import os 
 from flask import json
 from flask.wrappers import Request
 from colbert.infra import Run, RunConfig, ColBERTConfig
@@ -19,7 +20,7 @@ def index(self, request: Request):
 
         # collection = Collection(path="C:\\adverb-backend\\controllers\\downloads\\lotte\\science\\dev\\collection - Kopie.tsv")
 
-        checkpoint = "C:\\adverb-backend\\models\\colbertv2.0"
+        checkpoint = os.path.join(os.path.dirname(__file__), "..", "models", "colbertv2.0")
         nbits = 2   # encode each dimension with 2 bits
         doc_maxlen = 300   # truncate passages at 300 tokens
         with Run().context(RunConfig(nranks=4)):
@@ -44,4 +45,4 @@ def search_for_text(self, request: Request):
 
         # results = searcher.search(query, k=3)
 
-        # return results
+        return ""

run.bat

@@ -1,2 +1 @@
-call conda activate adverb
 set CUDA_VISIBLE_DEVICES=3 & python webservice.py

run.sh

@@ -1,4 +1,3 @@
 #!/bin/sh
-conda activate adverb
 CUDA_VISIBLE_DEVICES=3, python webservice.py 
 $SHELL

webservice.py

@@ -3,14 +3,17 @@
 os.environ["CUDA_VISIBLE_DEVICES"]="3"
 tf_device='/gpu:0'
 
+import sys
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), "src", "colbert", "colbert"))
+
 import argparse
 from flask import Flask, jsonify, request
 from flask_cors import CORS
 
 from controllers.code_summary_controller import CodeSummaryController
 from controllers.code_symbol_controller import CodeSymbolController
-from controllers.code_search_codebert_controller import CodeSearchCodeBertController
-# from controllers.code_search_colbert_controller import CodeSearchColBertController
+# from controllers.code_search_codebert_controller import CodeSearchCodeBertController
+from controllers.code_search_colbert_controller import CodeSearchColBertController
 
 DEBUG = True
 PORT = 8090
@@ -22,8 +25,8 @@
 # controllers
 code_summary = CodeSummaryController()
 code_symbol = CodeSymbolController()
-code_search_code_bert = CodeSearchCodeBertController()
-# code_search_col_bert = CodeSearchColBertController()
+# code_search_code_bert = CodeSearchCodeBertController()
+code_search_col_bert = CodeSearchColBertController()
 
 # API routes
 @app.route("/api/summary", methods = ["POST"])
@@ -34,8 +37,7 @@ def get_summary():
             response = jsonify(summary)
             response.headers.add("Access-Control-Allow-Origin", "*")
             return response
-    except Exception as e:
-        print(e)
+    except:
         pass
     return "Bad request", "400"
 
@@ -47,34 +49,33 @@ def get_symbol_name():
             response = jsonify(name)
             response.headers.add("Access-Control-Allow-Origin", "*")
             return response
-    except Exception as e:
-        print(e)
+    except:
         pass
     return "Bad request", "400"
 
 @app.route("/api/search", methods = ["POST"])
 def search():
     try:
-        search_result = code_search_code_bert.search_for_text(request)
-        # search_result = code_search_col_bert.search_for_text(request)
+        # search_result = code_search_code_bert.search_for_text(request)
+        search_result = code_search_col_bert.search_for_text(request)
         if search_result:
             response = jsonify(search_result)
             response.headers.add("Access-Control-Allow-Origin", "*")
             return response
     except Exception as e:
-        print(e)
+        print(str(e))
         pass
     return "Bad request", "400"
 
-# @app.route("/api/search_index", methods = ["POST"])
-# def search_index():
-#     try:
-#         code_search_col_bert.index(request)
-#         return "Success", "200"
-#     except Exception as e:
-#         print(e)
-#         pass
-#     return "Bad request", "400"
+@app.route("/api/search_index", methods = ["POST"])
+def search_index():
+    try:
+        code_search_col_bert.index(request)
+        return "Success", "200"
+    except Exception as e:
+        print(str(e))
+        pass
+    return "Bad request", "400"
 
 
 if __name__ == "__main__":