Hoja de Referencia de Generación de Código OpenAI Codex¶

Descripción General¶

OpenAI Codex es un potente sistema de IA que traduce lenguaje natural a código, capaz de comprender y generar código en docenas de lenguajes de programación. Construido sobre la arquitectura de GPT-3 y entrenado en miles de millones de líneas de código público, Codex impulsa GitHub Copilot y proporciona capacidades avanzadas de completación, generación y explicación de código. Destaca en comprender el contexto, generar funciones, clases y aplicaciones completas a partir de descripciones en lenguaje natural.

⚠️ Nota: La API de OpenAI Codex fue descontinuada en marzo de 2023. Esta guía cubre el uso histórico y la migración a los modelos GPT-3.5/GPT-4 para tareas de generación de código.

Migración a Modelos Actuales¶

GPT-3.5/GPT-4 para Generación de Código¶

# Install OpenAI Python library
pip install openai

# Basic setup for code generation with current models
import openai

openai.api_key = "your-api-key-here"

# Use GPT-3.5-turbo or GPT-4 for code generation
response = openai.ChatCompletion.create(
    model="gpt-3.5-turbo",  # or "gpt-4"
    messages=[
        \\\\{
            "role": "system",
            "content": "You are an expert programmer. Generate clean, efficient, and well-documented code."
        \\\\},
        \\\\{
            "role": "user",
            "content": "Write a Python function to implement binary search"
        \\\\}
    ],
    max_tokens=1000,
    temperature=0.1  # Lower temperature for more deterministic code
)

print(response.choices[0].message.content)

Uso de la API de Codex Legacy (Referencia Histórica)¶

# Historical Codex API usage (deprecated)
import openai

openai.api_key = "your-api-key-here"

# Codex completion (no longer available)
response = openai.Completion.create(
    engine="code-davinci-002",  # Deprecated
    prompt="# Function to calculate factorial\ndef factorial(n):",
    max_tokens=150,
    temperature=0,
    stop=["#", "\n\n"]
)

print(response.choices[0].text)

Configuración Moderna de Generación de Código¶

Configuración del SDK de Python¶

#!/usr/bin/env python3
# modern-codex-replacement.py

import openai
import os
import json
from typing import List, Dict, Optional
from datetime import datetime

class ModernCodeGenerator:
    def __init__(self, api_key: str = None):
        self.client = openai.OpenAI(
            api_key=api_key or os.getenv("OPENAI_API_KEY")
        )
        self.conversation_history = []

    def generate_code(self, prompt: str, language: str = "python",
                     model: str = "gpt-3.5-turbo") -> str:
        """Generate code using modern OpenAI models"""

        system_prompt = f"""
You are an expert \\\\{language\\\\} developer. Generate clean, efficient,
and well-documented code that follows best practices. Include:
- Proper error handling
- Type hints (where applicable)
- Comprehensive docstrings
- Security considerations
- Performance optimizations
"""

        try:
            response = self.client.chat.completions.create(
                model=model,
                messages=[
                    \\\\{"role": "system", "content": system_prompt\\\\},
                    \\\\{"role": "user", "content": prompt\\\\}
                ],
                max_tokens=2000,
                temperature=0.1,
                top_p=1.0,
                frequency_penalty=0.0,
                presence_penalty=0.0
            )

            generated_code = response.choices[0].message.content

            # Store in conversation history
            self.conversation_history.append(\\\\{
                "prompt": prompt,
                "language": language,
                "model": model,
                "response": generated_code,
                "timestamp": datetime.now().isoformat()
            \\\\})

            return generated_code

        except Exception as e:
            return f"Error generating code: \\\\{e\\\\}"

    def complete_code(self, partial_code: str, language: str = "python",
                     model: str = "gpt-3.5-turbo") -> str:
        """Complete partial code snippets"""

        prompt = f"""
Complete this \\\\{language\\\\} code snippet. Provide only the missing parts:

```{language}
{partial_code}

Continue the code logically and maintain the existing style and patterns. """

    try:
        response = self.client.chat.completions.create(
            model=model,
            messages=[
                \\\\{"role": "user", "content": prompt\\\\}
            ],
            max_tokens=1000,
            temperature=0.1
        )

        return response.choices[0].message.content

    except Exception as e:
        return f"Error completing code: \\\\{e\\\\}"

def explain_code(self, code: str, language: str = "python") -> str:
    """Explain existing code"""

    prompt = f"""

Explain this \\{language\\} code in detail:

{code}

Provide: 1. High-level overview 2. Line-by-line explanation of complex parts 3. Purpose and functionality 4. Potential improvements """

    try:
        response = self.client.chat.completions.create(
            model="gpt-3.5-turbo",
            messages=[
                \\\\{"role": "user", "content": prompt\\\\}
            ],
            max_tokens=1500,
            temperature=0.1
        )

        return response.choices[0].message.content

    except Exception as e:
        return f"Error explaining code: \\\\{e\\\\}"

def fix_code(self, buggy_code: str, error_message: str = None,
            language: str = "python") -> str:
    """Fix buggy code"""

    prompt = f"""

Fix this \\{language\\} code that has issues:

{buggy_code}

"""

    if error_message:
        prompt += f"\nError message: \\\\{error_message\\\\}"

    prompt += """

Provide: 1. Corrected code 2. Explanation of what was wrong 3. Prevention strategies for similar issues """

    try:
        response = self.client.chat.completions.create(
            model="gpt-3.5-turbo",
            messages=[
                \\\\{"role": "user", "content": prompt\\\\}
            ],
            max_tokens=1500,
            temperature=0.1
        )

        return response.choices[0].message.content

    except Exception as e:
        return f"Error fixing code: \\\\{e\\\\}"

def generate_tests(self, code: str, language: str = "python") -> str:
    """Generate test cases for code"""

    test_frameworks = \\\\{
        "python": "pytest",
        "javascript": "jest",
        "java": "junit",
        "csharp": "nunit",
        "go": "testing package"
    \\\\}

    framework = test_frameworks.get(language, "appropriate testing framework")

    prompt = f"""

Generate comprehensive test cases for this \\{language\\} code using \\{framework\\}:

{code}

Include: 1. Unit tests for all functions/methods 2. Edge cases and boundary conditions 3. Error handling tests 4. Mock objects where needed 5. Test data setup and teardown """

    try:
        response = self.client.chat.completions.create(
            model="gpt-3.5-turbo",
            messages=[
                \\\\{"role": "user", "content": prompt\\\\}
            ],
            max_tokens=2000,
            temperature=0.1
        )

        return response.choices[0].message.content

    except Exception as e:
        return f"Error generating tests: \\\\{e\\\\}"

Example usage¶

def main(): generator = ModernCodeGenerator()

# Generate a REST API client
prompt = """

Create a Python class for a REST API client that handles: - GET, POST, PUT, DELETE requests - Authentication with API keys - Retry logic with exponential backoff - Rate limiting - Response caching - Comprehensive error handling """

code = generator.generate_code(prompt, "python", "gpt-4")
print("Generated Code:")
print("=" * 50)
print(code)

if name == "main": main()

## Generación de Código Específico por Lenguaje

### Desarrollo en Python
```python
# Python-specific code generation examples

# Data science and machine learning
prompt = """
Create a Python script for machine learning that:
- Loads data from CSV files
- Performs data preprocessing and feature engineering
- Implements multiple ML models (Random Forest, SVM, Neural Network)
- Evaluates models with cross-validation
- Visualizes results and feature importance
- Saves the best model for deployment
"""

# Web development with FastAPI
prompt = """
Build a FastAPI application that:
- Implements user authentication with JWT
- Uses SQLAlchemy for database operations
- Includes CRUD operations for a blog system
- Implements rate limiting and CORS
- Includes comprehensive error handling
- Provides OpenAPI documentation
"""

# DevOps automation
prompt = """
Create a Python automation script that:
- Manages Docker containers
- Deploys applications to Kubernetes
- Monitors application health
- Sends alerts via Slack/email
- Implements rollback functionality
- Logs all operations
"""

Desarrollo en JavaScript/TypeScript¶

// JavaScript/TypeScript code generation

// React component with hooks
const prompt = `
Create a React TypeScript component that:
- Implements a data table with sorting, filtering, and pagination
- Uses React Query for data fetching
- Includes proper TypeScript interfaces
- Implements accessibility features (ARIA labels, keyboard navigation)
- Uses CSS modules for styling
- Includes comprehensive error boundaries
`;

// Node.js microservice
const prompt = `
Build a Node.js microservice with TypeScript that:
- Implements GraphQL API with Apollo Server
- Uses Prisma for database operations
- Includes authentication middleware
- Implements caching with Redis
- Uses Winston for logging
- Includes health check endpoints
`;

// Frontend build optimization
const prompt = `
Create a Webpack configuration that:
- Optimizes bundle size with code splitting
- Implements tree shaking
- Uses service workers for caching
- Includes source maps for debugging
- Supports hot module replacement
- Generates performance reports
`;

Desarrollo en Go¶

// Go-specific code generation

// Microservice with gRPC
prompt := `
Create a Go microservice that:
- Implements gRPC server with protocol buffers
- Uses GORM for database operations
- Includes middleware for logging and authentication
- Implements graceful shutdown
- Uses Prometheus for metrics
- Includes comprehensive error handling
`

// Concurrent data processing
prompt := `
Implement a Go program that:
- Processes large datasets concurrently
- Uses worker pools with goroutines
- Implements backpressure handling
- Includes progress monitoring
- Uses channels for communication
- Handles graceful cancellation
`

Técnicas Avanzadas de Generación de Código¶

Generación Consciente del Contexto¶

#!/usr/bin/env python3
# context-aware-generation.py

import openai
import os
import ast
from typing import List, Dict

class ContextAwareGenerator:
    def __init__(self):
        self.client = openai.OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
        self.project_context = \\\\{\\\\}

    def analyze_codebase(self, directory: str) -> Dict:
        """Analyze existing codebase for context"""

        context = \\\\{
            "languages": set(),
            "frameworks": set(),
            "patterns": set(),
            "dependencies": set(),
            "file_structure": \\\\{\\\\}
        \\\\}

        # Analyze Python files
        for root, dirs, files in os.walk(directory):
            for file in files:
                if file.endswith('.py'):
                    file_path = os.path.join(root, file)
                    try:
                        with open(file_path, 'r') as f:
                            content = f.read()

                        # Parse AST for imports and patterns
                        tree = ast.parse(content)

                        for node in ast.walk(tree):
                            if isinstance(node, ast.Import):
                                for alias in node.names:
                                    context["dependencies"].add(alias.name)
                            elif isinstance(node, ast.ImportFrom):
                                if node.module:
                                    context["dependencies"].add(node.module)

                        context["languages"].add("python")

                        # Detect frameworks
                        if "flask" in content.lower():
                            context["frameworks"].add("Flask")
                        if "django" in content.lower():
                            context["frameworks"].add("Django")
                        if "fastapi" in content.lower():
                            context["frameworks"].add("FastAPI")

                    except Exception as e:
                        print(f"Error analyzing \\\\{file_path\\\\}: \\\\{e\\\\}")

        self.project_context = context
        return context

    def generate_with_context(self, prompt: str, language: str = "python") -> str:
        """Generate code with project context"""

        context_info = ""
        if self.project_context:
            context_info = f"""
Project Context:
- Languages: \\\\{', '.join(self.project_context.get('languages', []))\\\\}
- Frameworks: \\\\{', '.join(self.project_context.get('frameworks', []))\\\\}
- Key Dependencies: \\\\{', '.join(list(self.project_context.get('dependencies', []))[:10])\\\\}

Please generate code that fits with this existing codebase.
"""

        full_prompt = f"\\\\{context_info\\\\}\n\nRequest: \\\\{prompt\\\\}"

        try:
            response = self.client.chat.completions.create(
                model="gpt-4",
                messages=[
                    \\\\{
                        "role": "system",
                        "content": f"You are an expert \\\\{language\\\\} developer working on an existing project. Generate code that integrates well with the existing codebase."
                    \\\\},
                    \\\\{"role": "user", "content": full_prompt\\\\}
                ],
                max_tokens=2000,
                temperature=0.1
            )

            return response.choices[0].message.content

        except Exception as e:
            return f"Error generating contextual code: \\\\{e\\\\}"

    def suggest_refactoring(self, code: str, language: str = "python") -> str:
        """Suggest refactoring based on project context"""

        context_info = ""
        if self.project_context:
            frameworks = ', '.join(self.project_context.get('frameworks', []))
            if frameworks:
                context_info = f"This project uses \\\\{frameworks\\\\}. "

        prompt = f"""
\\\\{context_info\\\\}Analyze this \\\\{language\\\\} code and suggest refactoring:

```{language}
{code}

Consider: 1. Code patterns used in the project 2. Framework-specific best practices 3. Performance optimizations 4. Maintainability improvements 5. Security enhancements """

    try:
        response = self.client.chat.completions.create(
            model="gpt-4",
            messages=[
                \\\\{"role": "user", "content": prompt\\\\}
            ],
            max_tokens=2000,
            temperature=0.1
        )

        return response.choices[0].message.content

    except Exception as e:
        return f"Error suggesting refactoring: \\\\{e\\\\}"

def main(): generator = ContextAwareGenerator()

# Analyze current project
context = generator.analyze_codebase(".")
print("Project Context:")
for key, value in context.items():
    print(f"  \\\\{key\\\\}: \\\\{value\\\\}")

# Generate code with context
prompt = "Create a new API endpoint for user management"
code = generator.generate_with_context(prompt)
print("\nGenerated Code:")
print("=" * 50)
print(code)

if name == "main": main()

### Generación de Código en Múltiples Pasos
```python
#!/usr/bin/env python3
# multi-step-generation.py

import openai
import os
from typing import List, Dict

class MultiStepGenerator:
    def __init__(self):
        self.client = openai.OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
        self.generation_steps = []

    def plan_implementation(self, requirement: str) -> List[str]:
        """Break down complex requirements into implementation steps"""

        prompt = f"""
Break down this software requirement into detailed implementation steps:

Requirement: \\\\{requirement\\\\}

Provide a step-by-step implementation plan with:
1. Architecture decisions
2. Component breakdown
3. Implementation order
4. Dependencies between components
5. Testing strategy

Format as a numbered list of specific, actionable steps.
"""

        try:
            response = self.client.chat.completions.create(
                model="gpt-4",
                messages=[
                    \\\\{"role": "user", "content": prompt\\\\}
                ],
                max_tokens=1500,
                temperature=0.1
            )

            plan = response.choices[0].message.content

            # Extract steps (simple parsing)
            steps = []
            for line in plan.split('\n'):
                if line.strip() and (line.strip()[0].isdigit() or line.strip().startswith('-')):
                    steps.append(line.strip())

            self.generation_steps = steps
            return steps

        except Exception as e:
            return [f"Error creating plan: \\\\{e\\\\}"]

    def implement_step(self, step: str, previous_code: str = "",
                      language: str = "python") -> str:
        """Implement a specific step"""

        context = ""
        if previous_code:
            context = f"""
Previous implementation:
```{language}
{previous_code}

Build upon this existing code. """

    prompt = f"""

\\{context\\}

Implement this specific step: \\{step\\}

Provide complete, working \\{language\\} code that: - Implements only this step - Integrates with previous code (if any) - Includes proper error handling - Follows best practices - Includes comments explaining the implementation """

    try:
        response = self.client.chat.completions.create(
            model="gpt-4",
            messages=[
                \\\\{"role": "user", "content": prompt\\\\}
            ],
            max_tokens=2000,
            temperature=0.1
        )

        return response.choices[0].message.content

    except Exception as e:
        return f"Error implementing step: \\\\{e\\\\}"

def generate_complete_solution(self, requirement: str,
                             language: str = "python") -> Dict:
    """Generate complete solution using multi-step approach"""

    print(f"Planning implementation for: \\\\{requirement\\\\}")

    # Step 1: Create implementation plan
    steps = self.plan_implementation(requirement)

    print(f"Implementation plan created with \\\\{len(steps)\\\\} steps")

    # Step 2: Implement each step
    complete_code = ""
    step_implementations = []

    for i, step in enumerate(steps, 1):
        print(f"Implementing step \\\\{i\\\\}/\\\\{len(steps)\\\\}: \\\\{step[:50]\\\\}...")

        step_code = self.implement_step(step, complete_code, language)
        step_implementations.append(\\\\{
            "step": step,
            "code": step_code,
            "step_number": i
        \\\\})

        # Accumulate code for next step
        complete_code += f"\n\n# Step \\\\{i\\\\}: \\\\{step\\\\}\n\\\\{step_code\\\\}"

    # Step 3: Review and optimize complete solution
    optimized_code = self.optimize_complete_solution(complete_code, language)

    return \\\\{
        "requirement": requirement,
        "language": language,
        "plan": steps,
        "step_implementations": step_implementations,
        "complete_code": complete_code,
        "optimized_code": optimized_code
    \\\\}

def optimize_complete_solution(self, code: str, language: str = "python") -> str:
    """Optimize the complete solution"""

    prompt = f"""

Review and optimize this complete \\{language\\} solution:

{code}

Optimize for: 1. Code organization and structure 2. Performance improvements 3. Error handling consistency 4. Code reuse and DRY principles 5. Documentation and comments 6. Security considerations

Provide the optimized version with explanations of changes. """

    try:
        response = self.client.chat.completions.create(
            model="gpt-4",
            messages=[
                \\\\{"role": "user", "content": prompt\\\\}
            ],
            max_tokens=3000,
            temperature=0.1
        )

        return response.choices[0].message.content

    except Exception as e:
        return f"Error optimizing solution: \\\\{e\\\\}"

def main(): generator = MultiStepGenerator()

requirement = """

Create a web scraping system that: - Scrapes product data from multiple e-commerce sites - Handles rate limiting and anti-bot measures - Stores data in a database - Provides a REST API to query the data - Includes monitoring and alerting - Supports distributed scraping across multiple workers """

solution = generator.generate_complete_solution(requirement, "python")

print("\nImplementation Plan:")
print("=" * 50)
for i, step in enumerate(solution["plan"], 1):
    print(f"\\\\{i\\\\}. \\\\{step\\\\}")

print(f"\nComplete solution generated with \\\\{len(solution['step_implementations'])\\\\} steps")
print("Check the generated files for detailed implementation.")

# Save results
with open("multi_step_solution.py", "w") as f:
    f.write(solution["optimized_code"])

print("Optimized solution saved to: multi_step_solution.py")

if name == "main": main()

## Integración con IDE y Editores

### Integración con VS Code
```json
// VS Code settings for modern code generation
\\\\{
  "openai.apiKey": "$\\\\{env:OPENAI_API_KEY\\\\}",
  "openai.model": "gpt-3.5-turbo",
  "openai.maxTokens": 1000,
  "openai.temperature": 0.1,
  "openai.codeCompletion": true,
  "openai.codeExplanation": true,
  "openai.codeGeneration": true,
  "openai.languages": [
    "python",
    "javascript",
    "typescript",
    "go",
    "java",
    "cpp",
    "csharp"
  ]
\\\\}

Plugin para Vim/Neovim¶

-- Neovim Lua configuration for code generation
local function generate_code()
  local prompt = vim.fn.input("Code generation prompt: ")
  if prompt == "" then
    return
  end

  local language = vim.bo.filetype
  local api_key = os.getenv("OPENAI_API_KEY")

  if not api_key then
    print("Error: OPENAI_API_KEY not set")
    return
  end

  -- Call Python script for code generation
  local cmd = string.format(
    "python3 -c \"import openai; openai.api_key='%s'; response=openai.ChatCompletion.create(model='gpt-3.5-turbo', messages=[\\\\{'role': 'user', 'content': 'Generate %s code: %s'\\\\}], max_tokens=1000, temperature=0.1); print(response.choices[0].message.content)\"",
    api_key, language, prompt
  )

  local output = vim.fn.system(cmd)

  -- Insert generated code at cursor
  local lines = vim.split(output, "\n")
  vim.api.nvim_put(lines, "l", true, true)
end

-- Key mapping
vim.keymap.set("n", "<leader>cg", generate_code, \\\\{ desc = "Generate code" \\\\})

Integración con Emacs¶

;; Emacs Lisp configuration for code generation
(defun openai-generate-code (prompt)
  "Generate code using OpenAI API"
  (interactive "sCode generation prompt: ")
  (let* ((api-key (getenv "OPENAI_API_KEY"))
         (language (file-name-extension (buffer-file-name)))
         (python-script (format
           "import openai; openai.api_key='%s'; response=openai.ChatCompletion.create(model='gpt-3.5-turbo', messages=[\\\\{'role': 'user', 'content': 'Generate %s code: %s'\\\\}], max_tokens=1000, temperature=0.1); print(response.choices[0].message.content)"
           api-key language prompt)))
    (if api-key
        (let ((output (shell-command-to-string (format "python3 -c \"%s\"" python-script))))
          (insert output))
      (message "Error: OPENAI_API_KEY not set"))))

;; Key binding
(global-set-key (kbd "C-c g") 'openai-generate-code)

Herramientas de Línea de Comandos¶

Generador de Código CLI¶

Would you like me to continue with the remaining placeholders?```bash

!/bin/bash¶

codegen-cli.sh - Command line code generator¶

OPENAI_API_KEY="$\\{OPENAI_API_KEY\\}" MODEL="gpt-3.5-turbo" MAX_TOKENS=1500 TEMPERATURE=0.1

if [ -z "$OPENAI_API_KEY" ]; then echo "Error: OPENAI_API_KEY environment variable not set" exit 1 fi

show_help() \\{ echo "Code Generation CLI Tool" echo "Usage:" echo " codegen-cli.sh generate - Generate code" echo " codegen-cli.sh complete - Complete code in file" echo " codegen-cli.sh explain - Explain code in file" echo " codegen-cli.sh fix [error_message] - Fix buggy code" echo " codegen-cli.sh test - Generate tests" echo "" echo "Examples:" echo " codegen-cli.sh generate python 'Create a REST API client'" echo " codegen-cli.sh complete main.py" echo " codegen-cli.sh explain algorithm.py" echo "" echo "Environment variables:" echo " OPENAI_API_KEY - Your OpenAI API key" \\}

call_openai_api() \\{ local system_prompt="$1" local user_prompt="$2"

python3 << EOF

import openai import json import sys

openai.api_key = "$OPENAI_API_KEY"

try: response = openai.ChatCompletion.create( model="$MODEL", messages=[ \\\\{"role": "system", "content": "$system_prompt"\\}, \\{"role": "user", "content": "$user_prompt"\\\\} ], max_tokens=$MAX_TOKENS, temperature=$TEMPERATURE )

print(response.choices[0].message.content)

except Exception as e: print(f"Error: \\{e\\}", file=sys.stderr) sys.exit(1) EOF \\}

case "$1" in "generate") if [ $# -lt 3 ]; then echo "Usage: codegen-cli.sh generate " exit 1 fi

    language="$2"
    prompt="$3"

    system_prompt="You are an expert $language developer. Generate clean, efficient, and well-documented code."
    user_prompt="Generate $language code: $prompt"

    echo "Generating $language code..."
    echo "=========================="
    call_openai_api "$system_prompt" "$user_prompt"
    ;;

"complete")
    if [ $# -lt 2 ]; then
        echo "Usage: codegen-cli.sh complete <file>"
        exit 1
    fi

    file="$2"

    if [ ! -f "$file" ]; then
        echo "Error: File $file not found"
        exit 1
    fi

    # Detect language from file extension
    extension="$\\\\{file##*.\\\\}"
    case "$extension" in
        "py") language="python" ;;
        "js") language="javascript" ;;
        "ts") language="typescript" ;;
        "go") language="go" ;;
        "java") language="java" ;;
        *) language="unknown" ;;
    esac

    code_content=$(cat "$file")
    user_prompt="Complete this $language code:\n\n$code_content"

    echo "Completing code in $file..."
    echo "=========================="
    call_openai_api "You are an expert programmer. Complete the provided code." "$user_prompt"
    ;;

"explain")
    if [ $# -lt 2 ]; then
        echo "Usage: codegen-cli.sh explain <file>"
        exit 1
    fi

    file="$2"

    if [ ! -f "$file" ]; then
        echo "Error: File $file not found"
        exit 1
    fi

    code_content=$(cat "$file")
    user_prompt="Explain this code in detail:\n\n$code_content"

    echo "Explaining code in $file..."
    echo "=========================="
    call_openai_api "You are an expert programmer. Explain code clearly and comprehensively." "$user_prompt"
    ;;

"fix")
    if [ $# -lt 2 ]; then
        echo "Usage: codegen-cli.sh fix <file> [error_message]"
        exit 1
    fi

    file="$2"
    error_message="$3"

    if [ ! -f "$file" ]; then
        echo "Error: File $file not found"
        exit 1
    fi

    code_content=$(cat "$file")
    user_prompt="Fix this buggy code:\n\n$code_content"

    if [ -n "$error_message" ]; then
        user_prompt="$user_prompt\n\nError message: $error_message"
    fi

    echo "Fixing code in $file..."
    echo "======================"
    call_openai_api "You are an expert debugger. Fix code issues and explain the problems." "$user_prompt"
    ;;

"test")
    if [ $# -lt 2 ]; then
        echo "Usage: codegen-cli.sh test <file>"
        exit 1
    fi

    file="$2"

    if [ ! -f "$file" ]; then
        echo "Error: File $file not found"
        exit 1
    fi

    code_content=$(cat "$file")
    user_prompt="Generate comprehensive test cases for this code:\n\n$code_content"

    echo "Generating tests for $file..."
    echo "============================="
    call_openai_api "You are an expert test engineer. Generate comprehensive test cases." "$user_prompt"
    ;;

*)
    show_help
    ;;

esac ## Mejores Prácticas y Optimizaciónpython

Best practices for code generation prompts¶

Specific and detailed prompts¶

good_prompt = """ Create a Python class for a database connection pool that: - Supports PostgreSQL and MySQL - Implements connection pooling with configurable min/max connections - Handles connection timeouts and retries - Includes health checks for connections - Provides async/await support - Implements proper resource cleanup - Includes comprehensive logging - Follows the context manager protocol """

Include constraints and requirements¶

constrained_prompt = """ Implement a rate limiter in Go that: - Uses the token bucket algorithm - Supports distributed rate limiting with Redis - Handles burst traffic up to 1000 requests/second - Includes metrics collection - Must be thread-safe - Should have minimal memory footprint - Include benchmark tests - Follow Go best practices and idioms """

Request specific output format¶

formatted_prompt = """ Generate a React TypeScript component with the following structure:

Interface definitions for props and state
Component implementation with hooks
CSS module styles
Unit tests with React Testing Library
Storybook stories for documentation

Component requirements: - Data table with sorting and filtering - Pagination support - Accessibility compliance (ARIA labels) - Responsive design - Error boundary handling """ ### Ingeniería de Prompts para Generación de Códigopython

!/usr/bin/env python3¶

code-quality-validator.py¶

import ast import subprocess import tempfile import os from typing import List, Dict

class CodeQualityValidator: def init(self): self.quality_checks = \\{ "python": self.validate_python_code, "javascript": self.validate_javascript_code, "typescript": self.validate_typescript_code \\}

def validate_generated_code(self, code: str, language: str) -> Dict:
    """Validate generated code quality"""

    validator = self.quality_checks.get(language)
    if not validator:
        return \\\\{"error": f"No validator for language: \\\\{language\\\\}"\\\\}

    return validator(code)

def validate_python_code(self, code: str) -> Dict:
    """Validate Python code quality"""

    issues = []

    # Syntax check
    try:
        ast.parse(code)
    except SyntaxError as e:
        issues.append(f"Syntax error: \\\\{e\\\\}")
        return \\\\{"valid": False, "issues": issues\\\\}

    # Create temporary file for linting
    with tempfile.NamedTemporaryFile(mode='w', suffix='.py', delete=False) as f:
        f.write(code)
        temp_file = f.name

    try:
        # Run flake8 for style checking
        result = subprocess.run(
            ['flake8', '--max-line-length=88', temp_file],
            capture_output=True,
            text=True
        )

        if result.returncode != 0:
            issues.extend(result.stdout.split('\n'))

        # Run bandit for security checking
        result = subprocess.run(
            ['bandit', '-f', 'txt', temp_file],
            capture_output=True,
            text=True
        )

        if result.returncode != 0 and "No issues identified" not in result.stdout:
            issues.append(f"Security issues found: \\\\{result.stdout\\\\}")

        # Check for best practices
        if 'import *' in code:
            issues.append("Avoid wildcard imports")

        if 'except:' in code and 'except Exception:' not in code:
            issues.append("Use specific exception handling")

        # Check for documentation
        if 'def ' in code and '"""' not in code:
            issues.append("Missing docstrings for functions")

    finally:
        os.unlink(temp_file)

    return \\\\{
        "valid": len(issues) == 0,
        "issues": [issue for issue in issues if issue.strip()],
        "language": "python"
    \\\\}

def validate_javascript_code(self, code: str) -> Dict:
    """Validate JavaScript code quality"""

    issues = []

    # Create temporary file
    with tempfile.NamedTemporaryFile(mode='w', suffix='.js', delete=False) as f:
        f.write(code)
        temp_file = f.name

    try:
        # Run ESLint if available
        result = subprocess.run(
            ['eslint', '--format', 'compact', temp_file],
            capture_output=True,
            text=True
        )

        if result.returncode != 0:
            issues.extend(result.stdout.split('\n'))

        # Basic checks
        if 'eval(' in code:
            issues.append("Avoid using eval() - security risk")

        if 'var ' in code:
            issues.append("Use 'let' or 'const' instead of 'var'")

    except FileNotFoundError:
        issues.append("ESLint not available - install for better validation")

    finally:
        os.unlink(temp_file)

    return \\\\{
        "valid": len(issues) == 0,
        "issues": [issue for issue in issues if issue.strip()],
        "language": "javascript"
    \\\\}

def validate_typescript_code(self, code: str) -> Dict:
    """Validate TypeScript code quality"""

    issues = []

    # Create temporary file
    with tempfile.NamedTemporaryFile(mode='w', suffix='.ts', delete=False) as f:
        f.write(code)
        temp_file = f.name

    try:
        # Run TypeScript compiler
        result = subprocess.run(
            ['tsc', '--noEmit', '--strict', temp_file],
            capture_output=True,
            text=True
        )

        if result.returncode != 0:
            issues.extend(result.stderr.split('\n'))

        # Check for TypeScript best practices
        if ': any' in code:
            issues.append("Avoid using 'any' type - use specific types")

        if '// @ts-ignore' in code:
            issues.append("Avoid @ts-ignore - fix type issues instead")

    except FileNotFoundError:
        issues.append("TypeScript compiler not available")

    finally:
        os.unlink(temp_file)

    return \\\\{
        "valid": len(issues) == 0,
        "issues": [issue for issue in issues if issue.strip()],
        "language": "typescript"
    \\\\}

def main(): validator = CodeQualityValidator()

# Example Python code validation
python_code = '''

def calculate_factorial(n): """Calculate factorial of a number.""" if n < 0: raise ValueError("Factorial not defined for negative numbers") if n == 0 or n == 1: return 1 return n * calculate_factorial(n - 1) '''

result = validator.validate_generated_code(python_code, "python")
print("Python Code Validation:")
print(f"Valid: \\\\{result['valid']\\\\}")
if result['issues']:
    print("Issues:")
    for issue in result['issues']:
        print(f"  - \\\\{issue\\\\}")

if name == "main": main() ### Validación de Calidad de Códigopython

Common migration issues from Codex to GPT models¶

Issue 1: Different response format¶

Old Codex format¶

response.choices[0].text¶

New GPT format¶

response.choices[0].message.content¶

Issue 2: Different prompt structure¶

Old Codex (completion)¶

prompt = "# Function to calculate factorial\ndef factorial(n):"

New GPT (chat)¶

messages = [ \\{"role": "system", "content": "You are an expert Python developer."\\}, \\{"role": "user", "content": "Write a function to calculate factorial"\\} ]

Issue 3: Stop sequences¶

Old Codex¶

stop=["#", "\n\n"]¶

New GPT (handled differently)¶

Use system prompts to control output format¶

## Resolución de Problemas e Issues Comunespython

Optimize API usage for better performance¶

class OptimizedCodeGenerator: def init(self): self.cache = \\{\\} self.batch_requests = []

def cached_generation(self, prompt: str, language: str) -> str:
    """Use caching to avoid duplicate requests"""

    cache_key = f"\\\\{language\\\\}:\\\\{hash(prompt)\\\\}"

    if cache_key in self.cache:
        return self.cache[cache_key]

    # Generate code
    result = self.generate_code(prompt, language)

    # Cache result
    self.cache[cache_key] = result

    return result

def batch_generation(self, prompts: List[str], language: str) -> List[str]:
    """Batch multiple requests for efficiency"""

    # Combine prompts
    combined_prompt = "Generate code for the following requests:\n\n"

    for i, prompt in enumerate(prompts, 1):
        combined_prompt += f"\\\\{i\\\\}. \\\\{prompt\\\\}\n\n"

    combined_prompt += f"Provide \\\\{language\\\\} code for each request, numbered accordingly."

    # Single API call
    response = self.generate_code(combined_prompt, language)

    # Parse responses (simplified)
    responses = response.split(f"\\\\{i+1\\\\}.")

    return responses[:len(prompts)]

```### Issues de Migración de API https://platform.openai.com/docs### Optimización de Rendimiento https://github.com/openai/openai-python## Recursos y Documentación https://platform.openai.com/docs/models/gpt-4### Recursos Oficiales - Documentación de OpenAI API - Librería Python de OpenAI - Documentación del Modelo GPT-4 - Cookbook de OpenAI https://community.openai.com/### Guías de Migración - Guía de Migración de Codex a GPT - Documentación de Migración de API - Mejores Prácticas para Generación de Código https://github.com/features/copilot### Recursos Comunitarios - Comunidad de Desarrolladores de OpenAI - Documentación de GitHub Copilot - Ejemplos de Generación de Código - [Guía de Ingeniería de Prompts](