Implement resume parsing :sweet:

.github/workflows/cd.yml

@@ -0,0 +1,49 @@
+name: CD
+
+on:
+  push:
+    branches: [main]
+
+env:
+  REGISTRY: ghcr.io
+  IMAGE_NAME: ${{ github.repository }}
+
+permissions:
+  packages: write
+
+jobs:
+  build:
+    name: Build Docker image
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+      - name: Install cosign
+        uses: sigstore/[email protected]
+        with:
+          cosign-release: v2.1.1
+      - name: Setup Docker buildx
+        uses: docker/[email protected]
+      - name: Log into registry ${{ env.REGISTRY }}
+        uses: docker/[email protected]
+        with:
+          registry: ${{ env.REGISTRY }}
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+      - name: Extract Docker metadata
+        id: meta
+        uses: docker/[email protected]
+        with:
+          images: "${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}"
+          tags: type=raw,value=latest
+      - name: Build & push Docker image
+        uses: docker/[email protected]
+        with:
+          context: .
+          platforms: linux/amd64,linux/arm64
+          push: "true"
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
+          cache-from: type=gha
+          cache-to: type=gha,mode=max=

README.md

@@ -2,15 +2,15 @@
 
 Dockerized microservice to extract information from PDF resumes of infinite kind and transform it into uniform JSON structure to be used for further automation.
 
-## Usage
+## Development
 
 Run below commands in project folder:
 
 ```shell
 # create app config
 cp config.dist.ini config.ini
 
-# update values in config.ini
+# update values in config.ini e.g., openai.* ones
 
 # start the services
 docker compose up -d

app/container.py

@@ -1,5 +1,6 @@
 from dependency_injector.containers import DeclarativeContainer, WiringConfiguration
-from dependency_injector.providers import Configuration
+from dependency_injector.providers import Configuration, Singleton
+from openai import OpenAI
 from os import path
 
 
@@ -13,3 +14,7 @@ class Container(DeclarativeContainer):
     wiring_config = WiringConfiguration(
         modules=[".routers.index"],
     )
+
+    openai = Singleton(
+        OpenAI, api_key=config.openai.api_key, organization=config.openai.organization
+    )

app/routers/index.py

@@ -1,9 +1,13 @@
 from dependency_injector.wiring import Provide, inject
-from fastapi import APIRouter, Depends, status
+from fastapi import APIRouter, Depends, File, HTTPException, UploadFile, status
+import json
+from openai import OpenAI
 from pydantic import BaseModel
-from typing import Any, Dict
+from pypdf import PdfReader
+from typing import Annotated, Any, Dict
 
 from app.container import Container
+from app.schema import Resume
 
 index_router = APIRouter()
 
@@ -18,3 +22,61 @@ async def index(
     config: Dict[str, Any] = Depends(Provide[Container.config]),
 ) -> IndexResponse:
     return IndexResponse(env=config["app"]["env"])
+
+
+@index_router.post(
+    "/process", status_code=status.HTTP_200_OK, response_model=Dict[str, Any]
+)
+@inject
+async def process(
+    resume: Annotated[UploadFile, File(description="Resume as a PDF file.")],
+    openai: OpenAI = Depends(Provide[Container.openai]),
+    config: Dict[str, Any] = Depends(Provide[Container.config]),
+) -> Dict[str, Any]:
+    if resume.content_type != "application/pdf":
+        raise HTTPException(
+            status_code=status.HTTP_400_BAD_REQUEST,
+            detail="Uploaded file is not a PDF file.",
+        )
+
+    # get the PDF content
+    pdf = PdfReader(resume.file)
+    num_pages = len(pdf.pages)
+    text = "\n".join(pdf.pages[page].extract_text() for page in range(num_pages))
+
+    # parse as JSON format
+    completion = openai.chat.completions.create(
+        messages=[
+            {
+                "role": "system",
+                "content": "You are a bot that parses information from resume's text content into structured JSON.",
+            },
+            {
+                "role": "user",
+                "content": "Following is the text content extracted from an uploaded PDF resume."
+                "Parse it into JSON format.",
+            },
+            {
+                "role": "user",
+                "content": text,
+            },
+        ],
+        model=config["openai"]["model"],
+        tools=[
+            {
+                "type": "function",
+                "function": {
+                    "name": "format_resume",
+                    "description": "Restructure data extracted from resume to the defined JSON schema.",
+                    "parameters": Resume,
+                },
+            }
+        ],
+        tool_choice={"type": "function", "function": {"name": "format_resume"}},
+    )
+
+    choice = completion.choices[0]
+    tool_call = choice.message.tool_calls[0]
+    output = json.loads(tool_call.function.arguments)
+
+    return output

app/schema.py

@@ -0,0 +1,119 @@
+Resume = {
+    "type": "object",
+    "properties": {
+        "name": {
+            "type": "string",
+            "description": "Full name of the candidate.",
+        },
+        "emails": {
+            "type": "array",
+            "description": "List of email addresses of the candidate.",
+            "items": {"type": "string"},
+        },
+        "phones": {
+            "type": "array",
+            "description": "List of phone numbers of the candidate, in E.164 format.",
+            "items": {"type": "string"},
+        },
+        "date_of_birth": {
+            "type": "string",
+            "description": "Date of birth of the candidate in YYYY-MM-DD format.",
+        },
+        "nationality": {
+            "type": "string",
+            "description": "Nationality of the candidate as ISO 3166-1 alpha-2 code.",
+        },
+        "about": {
+            "type": "string",
+            "description": "Text content from about section of the candidate.",
+        },
+        "skills": {
+            "type": "array",
+            "description": "List of skills mentioned by the candidate.",
+            "items": {"type": "string"},
+        },
+        "hobbies": {
+            "type": "array",
+            "description": "List of hobbies mentioned by the candidate.",
+            "items": {"type": "string"},
+        },
+        "languages": {
+            "type": "array",
+            "description": "List of languages known by the candidate as ISO 639 codes.",
+            "items": {"type": "string"},
+        },
+        "links": {
+            "type": "array",
+            "description": "List of URLs or links (website, social media etc.) mentioned by the candidate.",
+            "items": {"type": "string"},
+        },
+        "qualifications": {
+            "type": "array",
+            "description": "List of educational qualifications mentioned by the candidate sorted in descending order.",
+            "items": {
+                "type": "object",
+                "properties": {
+                    "degree": {
+                        "type": "string",
+                        "description": "Name of the degree which was achieved.",
+                    },
+                    "institution": {
+                        "type": "string",
+                        "description": "Name of the institution where this educational qualification was achieved.",
+                    },
+                    "started_at": {
+                        "type": "string",
+                        "description": "Start date of pursuing this educational qualification in YYYY-MM format.",
+                    },
+                    "ended_at": {
+                        "type": "string",
+                        "description": "End date of pursuing this educational qualification in YYYY-MM format."
+                        "Exclude if not mentioned of currently pursuing.",
+                    },
+                },
+                "required": ["degree", "institution"],
+            },
+        },
+        "experiences": {
+            "type": "array",
+            "description": "List of work experiences mentioned by the candidate sorted in descending order.",
+            "items": {
+                "type": "object",
+                "properties": {
+                    "company": {
+                        "type": "string",
+                        "description": "Name of the company or organization.",
+                    },
+                    "designation": {
+                        "type": "string",
+                        "description": "Name of the role or designation at this company.",
+                    },
+                    "location": {
+                        "type": "string",
+                        "description": "Location of the company.",
+                    },
+                    "description": {
+                        "type": "string",
+                        "description": "Job description, responsibilities or project details while working here.",
+                    },
+                    "started_at": {
+                        "type": "string",
+                        "description": "Joining date at this company in in YYYY-MM format.",
+                    },
+                    "exited_at": {
+                        "type": "string",
+                        "description": "Leaving or exit date at this company in YYYY-MM format."
+                        "Exclude if not mentioned of currently working.",
+                    },
+                    "skills": {
+                        "type": "array",
+                        "description": "List of skills learned or utilised while working here.",
+                        "items": {"type": "string"},
+                    },
+                },
+                "required": ["company", "designation", "from"],
+            },
+        },
+    },
+    "required": ["name"],
+}

app/web.py

@@ -1,8 +1,8 @@
 from fastapi import FastAPI
 from fastapi.middleware.cors import CORSMiddleware
 
-from .container import Container
-from .routers import index_router
+from app.container import Container
+from app.routers import index_router
 
 app = FastAPI()
 

config.dist.ini

@@ -1,2 +1,7 @@
 [app]
 env=dev
+
+[openai]
+api_key=
+organization=
+model=gpt-3.5-turbo