DeepSeek’s Latest Advances: A Strong Challenge to ChatGPT and Google

The field of artificial intelligence (AI) is experiencing intense competition… (rest of the original article content)
“””

rewritten_article = generate_text(prompt)

if rewritten_article:
print(rewritten_article)
else:
print(“Failed to generate article.”)


**Explanation:**

1.  **Import Necessary Libraries:** `requests` for making API calls and `json` for handling JSON data.
2.  **`generate_text(prompt)` Function:**
    *   Takes a `prompt` string as input. This prompt provides instructions to the AI.
    *   Sets up the API URL, Headers (including your API Key!), and Data payload.
    *   The `data` dictionary contains key-value pairs that tell the DeepSeek model *what* to do (the `prompt`), *which* model to use (`deepseek-r1-0528`), and *how* to do it (e.g., `max_tokens`, `temperature`).
    *   Uses `requests.post()` to send a POST request to the API.
    *   Handles potential errors during the API call (network issues, bad status codes) and when parsing the JSON response.  Crucially, it checks for `KeyError` if the expected fields aren't in the JSON. This is *very* important when working with APIs, as their output format can change.
    *   Returns the generated text, or `None` if there was an error.
3.  **Example Usage:**
    *   Defines a detailed and complex `prompt`. This prompt is the *key* to getting the desired output. The more specific you are, the better the results will be.
    *   Calls the `generate_text()` function with the `prompt`.
    *   Prints the returned article, or an error message if generation failed.
4.  **Crucial Improvements & Explanations:**
    *   **Error Handling:** The code now includes robust error handling to catch potential issues during the API call (network errors, invalid API keys, etc.) and provides informative error messages, including for missing fields.
        *   `response.raise_for_status()`:  This crucial line raises an HTTPError for bad responses (4xx or 5xx status codes).  Without this, you might silently get errors!
*   `except requests.exceptions.RequestException as e`: Catches a wide range of network-related errors.
        *   `except KeyError`:  Catches errors if the expected data fields ('choices', 'text') are not present in the API response. This is *essential* because APIs change!
    *   **API Key Security:**  *Never* hardcode your API key directly into your code, especially if you’re sharing it.  Instead, strongly consider setting up the API key in your environment variables and accessing it in python with `os.environ.get("DEEPSEEK_API_KEY")`. Replace `'Bearer YOUR_API_KEY'` with `f'Bearer {os.environ.get("DEEPSEEK_API_KEY")}'`
    *   **Prompt Engineering:** The prompt is very detailed.  The more specific and structured your prompt, the better the AI will be at following your instructions. This includes:
        *   Clearly stating the task: "Rewrite the following content..."
        *   Explicitly defining the constraints: "The title should not exceed..."
        *   Providing the desired output format: "Format the output as markdown code..."
        *   Specifying the target language: "in Turkish (Turkish)"
        *   Including the original content for reference.

**To run this code, you will need:**

1.  **A DeepSeek API Key:**  Obtain one from the DeepSeek website.
2.  **The `requests` Library:**  Install it using `pip install requests`.  Also, you might need the `python-dotenv` package if retrieving your key from environment variables.
3.  **Replace Placeholders:** Replace `'YOUR_API_KEY'` with your actual API key and ensure the `api_url` is valid.

This example demonstrates how to use DeepSeek R1-0528 for complex text generation tasks with specific formatting requirements.  Remember to adjust the `prompt` and other parameters to suit your specific use case.

güncellendi 2025-06-03

# LLM # AIGC # DeepSeek