Merge pull request Samagra-Development#254 from ksgr5566/search

Added word_score

Author: Gautam-Rajeev

config.json

@@ -78,6 +78,14 @@
       "environment": {},
       "nginx": []
     },
+    {
+      "serviceName": "word_score",
+      "modelBasePath": "src/search/word_score/local/.",
+      "apiBasePath": "/search/word_score/local",
+      "containerPort": 8000,
+      "environment": {},
+      "nginx": []
+    },
     {
       "serviceName": "text_translation_bhashini",
       "modelBasePath": "src/text_translation/bhashini/remote/.",

repository_data.json

@@ -120,6 +120,16 @@
                     "request_class": "ModelRequest"
                 }
             }
+        },
+        "search": {
+            "word_score": {
+                "local": {
+                    "__is_async": true,
+                     "__is_base": true,
+                    "model_class": "Model",
+                    "request_class": "ModelRequest"
+                }
+            }
         }
     }
 }

src/search/README.md

@@ -0,0 +1,3 @@
+# Purpose
+
+Common folder for scoring methods required for augmenting search and retrieval of documents.

src/search/__init__.py

@@ -0,0 +1 @@
+from word_score import *

src/search/word_score/README.md

@@ -0,0 +1,5 @@
+# Word Score
+
+This folder consists of an API that scores documents based on an approach that combines IDF and Fuzzy word matching.
+
+For a given query, it calculates fuzzy matching scores for words in query (max score for a word from entire row), weights them with IDF, takes average of the scores of all words in the query to give a score for the entire query, sorts them, and returns the top n matches.

src/search/word_score/__init__.py

@@ -0,0 +1 @@
+from .local import *

src/search/word_score/local/Dockerfile

@@ -0,0 +1,14 @@
+# Use an official Python runtime as a parent image
+FROM python:3.9-slim
+
+WORKDIR /app
+
+#install requirements
+COPY requirements.txt requirements.txt
+RUN pip3 install -r requirements.txt
+
+# Copy the rest of the application code to the working directory
+COPY . /app/
+EXPOSE 8000
+# Set the entrypoint for the container
+CMD ["hypercorn", "--bind", "0.0.0.0:8000", "api:app"]

src/search/word_score/local/README.md

@@ -0,0 +1,18 @@
+# Word Score
+
+## Test Deployment
+
+- Git clone the repo and cd to the project location.
+- cd to `local`, i.e., `cd ./src/search/word_score/local`.
+- Replace the file in `./content` with a csv file of your choice, but the data column should be named `tags` column.
+- Start your docker engine and `docker build -t word_score .`.
+- Do `docker run -p 8000:8000 word_score`.
+- `curl -X POST -H "Content-Type: application/json" -d '{"query": QUERY, "n": N}' http://0.0.0.0:8000`. <br> Replace `QUERY` with a query and `N` with the number of rows you want to retrieve.
+- The reponse for above would be: <br>
+`
+{
+    "docs": ["row1", "row2", ... , "rowN"]
+}
+`
+The list of strings contains the top N rows.
+

src/search/word_score/local/__init__.py

@@ -0,0 +1,2 @@
+from .request import *
+from .model import *

src/search/word_score/local/api.py

@@ -0,0 +1,26 @@
+from model import Model
+from request import ModelRequest
+from quart import Quart, request
+import aiohttp
+
+#from fastapi import FastAPI, Body
+app = Quart(__name__)
+#app.client = aiohttp.ClientSession()
+#app = FastAPI()
+
+@app.before_serving
+async def startup():
+    app.client = aiohttp.ClientSession()
+
+@app.route('/', methods=['POST'])
+async def translate():
+    data = await request.get_json()
+    req = ModelRequest(**data)
+    model = Model(app)
+    return await model.inference(req)
+
+@app.route('/', methods=['GET'])
+async def hi():
+    return "hi"
+
+