Error: fetch_error: fetch failed

[shanbady]

Getting:

Error: fetch_error: fetch failed
    at generate (file:///usr/local/lib/node_modules/codespin/dist/commands/generate.js:118:19)
    at async Object.handler (file:///usr/local/lib/node_modules/codespin/dist/index.js:117:9)

when using:
codespin gen --prompt '<some promt>'

my openai key is valid and set. any ideas?

[jeswin]

Hey @shanbady,

Can you let me know which version of codespin you're using?

You can get this with codespin version.

[shanbady]

latest i think. I installed via npm install -g codespin

[shanbady]

codespin --version says 0.0.62

[jeswin]

Can you post please:

• the full command (including the prompt)
• Have you done codespin init? This is not required, but just trying to understand your environment.

Also, I can send you Google Meet link to do some debugging. :)

[jeswin]

Can you check if api.openai.com is accessible from your computer?

[shanbady]

It is - when i hit it it responds with { "error": { "message": "Invalid URL (GET /)", "type": "invalid_request_error", "param": null, "code": null } }

[jeswin]

Please share:

• the full command (including the prompt)
• Have you done codespin init? This is not required, but just trying to understand your environment.

I am not able to reproduce at my end; so we'll need to dig deeper.

[shanbady]

I tried again with the vpn off. still getting the same.

[jeswin]

I just tried:
codespin gen --prompt 'Create a typescript function that adds two numbers'

And received:

FILE: addNumbers.ts
-------------------
function addNumbers(a: number, b: number): number {
  return a + b;
}

[jeswin]

Where are you located?

[shanbady]

i've done a codespin init (have also tried without.

full prompt:

codespin gen --prompt 'Create a django app for email marketing to customers. This app should be multi-tenant and support multiple organizations. A user should be able to sign up for an an account, create an organization and then create a workspace. They should then be able to import customers either through csv or using the api. The app should have an "events" api where a customer could send arbitrary json formatted events associated with a customer id in the database (it should add a new customer if one does not exist). There should be an interface for viewing all event data, a view for viewing all contacts and their data. The user should also have the ability to create "segments" based on customer attribute conditions(for example, a segment could be "all contacts who have an '@gmail' email address, who have a 'birth-date' after 1999 and have a first_name attribute of 'Fred'). The main application database should be postgres. Stream processing and contact segmentation should be done using ksqlb or a no-sql database such as mongodb. Nginx should be used as the server/load balancer. The entire application should be dockerized using docker-compose.' --write

[shanbady]

node -v is v19.6.0

npm -v is 9.6.7

[jeswin]

Thank you.

I have a suspicion it's not able to hit the API. There are similar issues reported here: transitive-bullshit/agentic#419

That's specific to China. If you're in Boston (guessing from your profile), that wouldn't be the reason. Unless you're on a VPN, or tor etc.

[jeswin]

Let me see if fetch exists in 19.6. I should have thought of that.

Add:
Tested. Exists.

[jeswin]

Ah wait. There's a better way to do this.

Can you append the --debug flag to the command please?

[jeswin]

You can use the following command:
codespin gen --prompt 'Create a typescript function that adds two numbers'

Once we figure it out with the prompt above, we can move to the larger prompt.

[shanbady]

--- PROMPT ---
Create a typescript function that adds two numbers

-----

Respond with just the code (but exclude invocation examples etc) in the following format:

$START_FILE_CONTENTS:./some/path/filename.ext$
import a from "./a";
function somethingSomething() {
  //....
}
$END_FILE_CONTENTS:./some/path/filename.ext$


OPENAI: model=gpt-3.5-turbo
OPENAI: maxTokens=3693
---OPENAI RESPONSE---
{"data":{"id":"chatcmpl-8AsA3Ta27vbLvhjcxRUbp0pwERYDD","object":"chat.completion","created":1697602167,"model":"gpt-3.5-turbo-0613","choices":[{"index":0,"message":{"role":"assistant","content":"$START_FILE_CONTENTS:addNumbers.ts$\nfunction addNumbers(a: number, b: number): number {\n  return a + b;\n}\n$END_FILE_CONTENTS:addNumbers.ts$"},"finish_reason":"stop"}],"usage":{"prompt_tokens":76,"completion_tokens":38,"total_tokens":114}}}
FILE: addNumbers.ts
-------------------
function addNumbers(a: number, b: number): number {
  return a + b;
}

[jeswin]

Ah, so that worked.

I tried your prompt, and it kinda worked as well.

FILE: ./docker-compose.yml
--------------------------
version: '3'

services:
  nginx:
    image: nginx:latest
    ports:
      - 80:80
    volumes:
      - ./nginx.conf:/etc/nginx/nginx.conf
    depends_on:
      - web

  web:
    build:
      context: .
      dockerfile: Dockerfile
    command: python manage.py runserver 0.0.0.0:8000
    volumes:
      - .:/code
    ports:
      - 8000:8000
    depends_on:
      - db

  db:
    image: postgres:latest
    environment:
      - POSTGRES_USER=postgres
      - POSTGRES_PASSWORD=postgres
      - POSTGRES_DB=email_marketing

FILE: ./Dockerfile
------------------
FROM python:3.9

ENV PYTHONUNBUFFERED 1

WORKDIR /code

COPY requirements.txt /code/
RUN pip install -r requirements.txt

COPY . /code/

FILE: ./requirements.txt
------------------------
Django==3.2.7
psycopg2-binary==2.9.1

FILE: ./email_marketing/settings.py
-----------------------------------
import os

# ...

DATABASES = {
    'default': {
        'ENGINE': 'django.db.backends.postgresql',
        'NAME': os.environ.get('DB_NAME', 'email_marketing'),
        'USER': os.environ.get('DB_USER', 'postgres'),
        'PASSWORD': os.environ.get('DB_PASSWORD', 'postgres'),
        'HOST': os.environ.get('DB_HOST', 'db'),
        'PORT': os.environ.get('DB_PORT', '5432'),
    }
}

# ...

FILE: ./email_marketing/urls.py
-------------------------------
from django.contrib import admin
from django.urls import path, include

urlpatterns = [
    path('admin/', admin.site.urls),
    path('api/', include('api.urls')),
]

FILE: ./api/urls.py
-------------------
from django.urls import path
from . import views

urlpatterns = [
    path('signup/', views.signup, name='signup'),
    path('organizations/', views.OrganizationListCreateView.as_view(), name='organization-list-create'),
    path('workspaces/', views.WorkspaceListCreateView.as_view(), name='workspace-list-create'),
    path('import/customers/', views.import_customers, name='import-customers'),
    path('events/', views.EventListCreateView.as_view(), name='event-list-create'),
    path('contacts/', views.ContactListCreateView.as_view(), name='contact-list-create'),
    path('segments/', views.SegmentListCreateView.as_view(), name='segment-list-create'),
]

FILE: ./api/views.py
--------------------
from django.contrib.auth.models import User
from django.shortcuts import render
from rest_framework import generics, status
from rest_framework.decorators import api_view
from rest_framework.response import Response
from .models import Organization, Workspace, Customer, Event, Contact, Segment
from .serializers import (
    UserSerializer, OrganizationSerializer, WorkspaceSerializer,
    CustomerSerializer, EventSerializer, ContactSerializer, SegmentSerializer
)

@api_view(['POST'])
def signup(request):
    serializer = UserSerializer(data=request.data)
    if serializer.is_valid():
        user = serializer.save()
        return Response(serializer.data, status=status.HTTP_201_CREATED)
    return Response(serializer.errors, status=status.HTTP_400_BAD_REQUEST)

class OrganizationListCreateView(generics.ListCreateAPIView):
    queryset = Organization.objects.all()
    serializer_class = OrganizationSerializer

class WorkspaceListCreateView(generics.ListCreateAPIView):
    queryset = Workspace.objects.all()
    serializer_class = WorkspaceSerializer

@api_view(['POST'])
def import_customers(request):
    # Import customers logic
    return Response(status=status.HTTP_200_OK)

class EventListCreateView(generics.ListCreateAPIView):
    queryset = Event.objects.all()
    serializer_class = EventSerializer

class ContactListCreateView(generics.ListCreateAPIView):
    queryset = Contact.objects.all()
    serializer_class = ContactSerializer

class SegmentListCreateView(generics.ListCreateAPIView):
    queryset = Segment.objects.all()
    serializer_class = SegmentSerializer

FILE: ./api/models.py
---------------------
from django.db import models

class Organization(models.Model):
    name = models.CharField(max_length=255)

    def __str__(self):
        return self.name

class Workspace(models.Model):
    name = models.CharField(max_length=255)
    organization = models.ForeignKey(Organization, on_delete=models.CASCADE)

    def __str__(self):
        return self.name

class Customer(models.Model):
    # Customer fields
    workspace = models.ForeignKey(Workspace, on_delete=models.CASCADE)

    def __str__(self):
        return f"Customer {self.id}"

class Event(models.Model):
    customer = models.ForeignKey(Customer, on_delete=models.CASCADE)
    data = models.JSONField()

class Contact(models.Model):
    # Contact fields
    workspace = models.ForeignKey(Workspace, on_delete=models.CASCADE)

    def __str__(self):
        return f"Contact {self.id}"

class Segment(models.Model):
    name = models.CharField(max_length=255)
    workspace = models.ForeignKey(Workspace, on_delete=models.CASCADE)
    conditions = models.JSONField()

    def __str__(self):
        return self.name

FILE: ./api/serializers.py
--------------------------
from django.contrib.auth.models import User
from rest_framework import serializers
from .models import Organization, Workspace, Customer, Event, Contact, Segment

class UserSerializer(serializers.ModelSerializer):
    class Meta:
        model = User
        fields = ['username', 'password']
        extra_kwargs = {'password': {'write_only': True}}

    def create(self, validated_data):
        user = User.objects.create_user(**validated_data)
        return user

class OrganizationSerializer(serializers.ModelSerializer):
    class Meta:
        model = Organization
        fields = '__all__'

class WorkspaceSerializer(serializers.ModelSerializer):
    class Meta:
        model = Workspace
        fields = '__all__'

class CustomerSerializer(serializers.ModelSerializer):
    class Meta:
        model = Customer
        fields = '__all__'

class EventSerializer(serializers.ModelSerializer):
    class Meta:
        model = Event
        fields = '__all__'

class ContactSerializer(serializers.ModelSerializer):
    class Meta:
        model = Contact
        fields = '__all__'

class SegmentSerializer(serializers.ModelSerializer):
    class Meta:
        model = Segment
        fields = '__all__'

FILE: ./nginx.conf
------------------
worker_processes 1;

events { worker_connections 1024; }

http {
    sendfile on;

    upstream web {
        server web:8000;
    }

    server {
        listen 80;

        location / {
            proxy_pass http://web;
            proxy_set_header Host $host;
            proxy_set_header X-Real-IP $remote_addr;
            proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        }
    }
}

[jeswin]

Some general points though:

1. For more advanced use cases, think of a file structure for your project, and define the prompts individually for each file.
2. You're at the limit of exceeding default token sizes. You might to change max-tokens when generating multiple files.
3. But the default model doesn't support large context sizes. You might need to switch to "gpt-3.5-turbo-16k" or "gpt-4" (GPT4 being way more expensive), and then specify max-tokens.

[jeswin]

@shanbady Can you verify, so that I can close this?

[shanbady]

That seems to have been it. you can close this