Add support for OpenAI-compatible API providers (e.g., MiniMax)

README.md

@@ -147,14 +147,16 @@ You can follow these steps to generate a PageIndex tree from a PDF document.
 pip3 install --upgrade -r requirements.txt
 ```
 
-### 2. Set your OpenAI API key
+### 2. Set your API key
 
 Create a `.env` file in the root directory and add your API key:
 
 ```bash
 CHATGPT_API_KEY=your_openai_key_here
 ```
 
+You can also use `OPENAI_API_KEY` as an alternative environment variable name.
+
 ### 3. Run PageIndex on your PDF
 
 ```bash
@@ -167,7 +169,9 @@ python3 run_pageindex.py --pdf_path /path/to/your/document.pdf
 You can customize the processing with additional optional arguments:
 
 ```
---model                 OpenAI model to use (default: gpt-4o-2024-11-20)
+--model                 Model to use (default: gpt-4o-2024-11-20)
+--api-key               API key (overrides env var)
+--base-url              Base URL for OpenAI-compatible API providers
 --toc-check-pages       Pages to check for table of contents (default: 20)
 --max-pages-per-node    Max pages per node (default: 10)
 --max-tokens-per-node   Max tokens per node (default: 20000)
@@ -177,6 +181,39 @@ You can customize the processing with additional optional arguments:
 ```
 </details>
 
+<details>
+<summary><strong>Using alternative LLM providers</strong></summary>
+<br>
+
+PageIndex supports any OpenAI-compatible API provider via the `--base-url` flag or the `OPENAI_BASE_URL` environment variable.
+
+**MiniMax**
+
+[MiniMax](https://www.minimaxi.com) offers models like `MiniMax-M2.5` with 204K context window through an OpenAI-compatible API:
+
+```bash
+# Via environment variables
+OPENAI_API_KEY=your_minimax_key_here
+OPENAI_BASE_URL=https://api.minimax.io/v1
+
+python3 run_pageindex.py --pdf_path /path/to/your/document.pdf --model MiniMax-M2.5
+```
+
+Or via CLI arguments:
+
+```bash
+python3 run_pageindex.py --pdf_path /path/to/your/document.pdf \
+  --model MiniMax-M2.5 \
+  --api-key your_minimax_key_here \
+  --base-url https://api.minimax.io/v1
+```
+
+**Other compatible providers**
+
+Any provider offering an OpenAI-compatible chat completions API can be used in the same way by setting `--base-url` to their endpoint.
+
+</details>
+
 <details>
 <summary><strong>Markdown support</strong></summary>
 <br>

pageindex/utils.py

@@ -17,7 +17,8 @@
 from pathlib import Path
 from types import SimpleNamespace as config
 
-CHATGPT_API_KEY = os.getenv("CHATGPT_API_KEY")
+CHATGPT_API_KEY = os.getenv("CHATGPT_API_KEY") or os.getenv("OPENAI_API_KEY")
+OPENAI_BASE_URL = os.getenv("OPENAI_BASE_URL")
 
 def count_tokens(text, model=None):
     if not text:
@@ -26,21 +27,24 @@ def count_tokens(text, model=None):
     tokens = enc.encode(text)
     return len(tokens)
 
-def ChatGPT_API_with_finish_reason(model, prompt, api_key=CHATGPT_API_KEY, chat_history=None):
+def ChatGPT_API_with_finish_reason(model, prompt, api_key=CHATGPT_API_KEY, chat_history=None, base_url=OPENAI_BASE_URL):
     max_retries = 10
-    client = openai.OpenAI(api_key=api_key)
+    client_kwargs = {"api_key": api_key}
+    if base_url:
+        client_kwargs["base_url"] = base_url
+    client = openai.OpenAI(**client_kwargs)
     for i in range(max_retries):
         try:
             if chat_history:
                 messages = chat_history
                 messages.append({"role": "user", "content": prompt})
             else:
                 messages = [{"role": "user", "content": prompt}]
-            
+
             response = client.chat.completions.create(
                 model=model,
                 messages=messages,
-                temperature=0,
+                temperature=0.01 if base_url else 0,
             )
             if response.choices[0].finish_reason == "length":
                 return response.choices[0].message.content, "max_output_reached"
@@ -58,21 +62,24 @@ def ChatGPT_API_with_finish_reason(model, prompt, api_key=CHATGPT_API_KEY, chat_
 
 
 
-def ChatGPT_API(model, prompt, api_key=CHATGPT_API_KEY, chat_history=None):
+def ChatGPT_API(model, prompt, api_key=CHATGPT_API_KEY, chat_history=None, base_url=OPENAI_BASE_URL):
     max_retries = 10
-    client = openai.OpenAI(api_key=api_key)
+    client_kwargs = {"api_key": api_key}
+    if base_url:
+        client_kwargs["base_url"] = base_url
+    client = openai.OpenAI(**client_kwargs)
     for i in range(max_retries):
         try:
             if chat_history:
                 messages = chat_history
                 messages.append({"role": "user", "content": prompt})
             else:
                 messages = [{"role": "user", "content": prompt}]
-            
+
             response = client.chat.completions.create(
                 model=model,
                 messages=messages,
-                temperature=0,
+                temperature=0.01 if base_url else 0,
             )
 
             return response.choices[0].message.content
@@ -86,16 +93,19 @@ def ChatGPT_API(model, prompt, api_key=CHATGPT_API_KEY, chat_history=None):
                 return "Error"
 
 
-async def ChatGPT_API_async(model, prompt, api_key=CHATGPT_API_KEY):
+async def ChatGPT_API_async(model, prompt, api_key=CHATGPT_API_KEY, base_url=OPENAI_BASE_URL):
     max_retries = 10
     messages = [{"role": "user", "content": prompt}]
+    client_kwargs = {"api_key": api_key}
+    if base_url:
+        client_kwargs["base_url"] = base_url
     for i in range(max_retries):
         try:
-            async with openai.AsyncOpenAI(api_key=api_key) as client:
+            async with openai.AsyncOpenAI(**client_kwargs) as client:
                 response = await client.chat.completions.create(
                     model=model,
                     messages=messages,
-                    temperature=0,
+                    temperature=0.01 if base_url else 0,
                 )
                 return response.choices[0].message.content
         except Exception as e:

run_pageindex.py

@@ -11,6 +11,10 @@
     parser.add_argument('--md_path', type=str, help='Path to the Markdown file')
 
     parser.add_argument('--model', type=str, default='gpt-4o-2024-11-20', help='Model to use')
+    parser.add_argument('--api-key', type=str, default=None,
+                      help='API key (defaults to CHATGPT_API_KEY or OPENAI_API_KEY env var)')
+    parser.add_argument('--base-url', type=str, default=None,
+                      help='Base URL for OpenAI-compatible API (e.g., https://api.minimax.io/v1)')
 
     parser.add_argument('--toc-check-pages', type=int, default=20, 
                       help='Number of pages to check for table of contents (PDF only)')
@@ -36,7 +40,13 @@
     parser.add_argument('--summary-token-threshold', type=int, default=200,
                       help='Token threshold for generating summaries (markdown only)')
     args = parser.parse_args()
-
+
+    # Override env vars if CLI args are provided
+    if args.api_key:
+        os.environ["CHATGPT_API_KEY"] = args.api_key
+    if args.base_url:
+        os.environ["OPENAI_BASE_URL"] = args.base_url
+
     # Validate that exactly one file type is specified
     if not args.pdf_path and not args.md_path:
         raise ValueError("Either --pdf_path or --md_path must be specified")