Add support for new LLMs and update tests

README.md

@@ -84,19 +84,21 @@ Based on several experiments, SlideDeck AI generally recommends the use of **Mis
 
 The supported LLMs offer different styles of content generation. Use one of the following LLMs along with relevant API keys/access tokens, as appropriate, to create the content of the slide deck:
 
-| LLM                              | Provider (code) | Requires API key                                                                                                         | Characteristics          |
-|:---------------------------------| :------- |:-------------------------------------------------------------------------------------------------------------------------|:-------------------------|
-| Gemini 2.0 Flash                 | Google Gemini API (`gg`) | Mandatory; [get here](https://aistudio.google.com/apikey)                                                                | Faster, longer content   |
-| Gemini 2.0 Flash Lite            | Google Gemini API (`gg`) | Mandatory; [get here](https://aistudio.google.com/apikey)                                                                | Fastest, longer content  |
-| Gemini 2.5 Flash                 | Google Gemini API (`gg`) | Mandatory; [get here](https://aistudio.google.com/apikey)                                                                | Faster, longer content   |
-| Gemini 2.5 Flash Lite            | Google Gemini API (`gg`) | Mandatory; [get here](https://aistudio.google.com/apikey)                                                                | Fastest, longer content  |
-| GPT                              | Azure OpenAI (`az`)      | Mandatory; [get here](https://ai.azure.com/resource/playground)  NOTE: You need to have your subscription/billing set up | Faster, longer content   |
-| Command R+                       | Cohere (`co`) | Mandatory; [get here](https://dashboard.cohere.com/api-keys)                                                             | Shorter, simpler content |
-| Gemini-2.0-flash-001             | OpenRouter (`or`) | Mandatory; [get here](https://openrouter.ai/settings/keys)                                                                            | Faster, longer content   |
-| GPT-3.5 Turbo                    | OpenRouter (`or`) | Mandatory; [get here](https://openrouter.ai/settings/keys)                                                                            | Faster, longer content   |
-| DeepSeek V3-0324                 | Together AI (`to`) | Mandatory; [get here](https://api.together.ai/settings/api-keys)                                                         | Slower, medium-length    |
-| Llama 3.3 70B Instruct Turbo     | Together AI (`to`) | Mandatory; [get here](https://api.together.ai/settings/api-keys)                                                         | Slower, detailed         |
-| Llama 3.1 8B Instruct Turbo 128K | Together AI (`to`) | Mandatory; [get here](https://api.together.ai/settings/api-keys)                                                         | Faster, shorter          |
+| LLM                                 | Provider (code)                                                                                                                                    | Requires API key                                                                                                         | Characteristics          |
+|:------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------------------------------------------|:-------------------------|
+| Claude Haiku 4.5                    |  Anthropic (`an`)                                                                                                                                  | Mandatory; [get here](https://platform.claude.com/settings/keys)                                                         | Faster, detailed         |
+| Gemini 2.0 Flash                    | Google Gemini API (`gg`)                                                                                                                           | Mandatory; [get here](https://aistudio.google.com/apikey)                                                                | Faster, longer content   |
+| Gemini 2.0 Flash Lite               | Google Gemini API (`gg`)                                                                                                                           | Mandatory; [get here](https://aistudio.google.com/apikey)                                                                | Fastest, longer content  |
+| Gemini 2.5 Flash                    | Google Gemini API (`gg`)                                                                                                                           | Mandatory; [get here](https://aistudio.google.com/apikey)                                                                | Faster, longer content   |
+| Gemini 2.5 Flash Lite               | Google Gemini API (`gg`)                                                                                                                           | Mandatory; [get here](https://aistudio.google.com/apikey)                                                                | Fastest, longer content  |
+| GPT                                 | Azure OpenAI (`az`)                                                                                                                                | Mandatory; [get here](https://ai.azure.com/resource/playground)  NOTE: You need to have your subscription/billing set up | Faster, longer content   |
+| Command R+                          | Cohere (`co`)                                                                                                                                      | Mandatory; [get here](https://dashboard.cohere.com/api-keys)                                                             | Shorter, simpler content |
+| Gemini-2.0-flash-001                | OpenRouter (`or`)                                                                                                                                  | Mandatory; [get here](https://openrouter.ai/settings/keys)                                                               | Faster, longer content   |
+| GPT-3.5 Turbo                       | OpenRouter (`or`)                                                                                                                                  | Mandatory; [get here](https://openrouter.ai/settings/keys)                                                               | Faster, longer content   |
+| DeepSeek V3-0324                    | Together AI (`to`)                                                                                                                                 | Mandatory; [get here](https://api.together.ai/settings/api-keys)                                                         | Slower, medium-length    |
+| Llama 3.3 70B Instruct Turbo        | Together AI (`to`)                                                                                                                                 | Mandatory; [get here](https://api.together.ai/settings/api-keys)                                                         | Slower, detailed         |
+| Llama 3.1 8B Instruct Turbo 128K    | Together AI (`to`)                                                                                                                                 | Mandatory; [get here](https://api.together.ai/settings/api-keys)                                                         | Faster, shorter          |
+| Llama-3.3-Swallow-70B-Instruct-v0.4 | SambaNova (`sn`)                                                                                                                                   | Mandatory; [get here](https://cloud.sambanova.ai/apis)                                                                   | Fast, shorter            |
 
 > **IMPORTANT**: SlideDeck AI does **NOT** store your API keys/tokens or transmit them elsewhere. If you provide your API key, it is only used to invoke the relevant LLM to generate contents. That's it! This is an 
 Open-Source project, so feel free to audit the code and convince yourself. 

src/slidedeckai/global_config.py

@@ -19,33 +19,52 @@ class GlobalConfig:
     """
     A data class holding the configurations.
     """
-
+    PROVIDER_ANTHROPIC = 'an'
+    PROVIDER_AZURE_OPENAI = 'az'
     PROVIDER_COHERE = 'co'
     PROVIDER_GOOGLE_GEMINI = 'gg'
-    PROVIDER_HUGGING_FACE = 'hf'
-    PROVIDER_AZURE_OPENAI = 'az'
     PROVIDER_OLLAMA = 'ol'
     PROVIDER_OPENROUTER = 'or'
     PROVIDER_TOGETHER_AI = 'to'
+    PROVIDER_SAMBANOVA = 'sn'
+
+    LITELLM_PROVIDER_MAPPING = {
+        PROVIDER_ANTHROPIC: 'anthropic',
+        PROVIDER_GOOGLE_GEMINI: 'gemini',
+        PROVIDER_AZURE_OPENAI: 'azure',
+        PROVIDER_OPENROUTER: 'openrouter',
+        PROVIDER_COHERE: 'cohere',
+        PROVIDER_SAMBANOVA: 'sambanova',
+        PROVIDER_TOGETHER_AI: 'together_ai',
+        PROVIDER_OLLAMA: 'ollama',
+    }
+
     VALID_PROVIDERS = {
+        PROVIDER_ANTHROPIC,
+        PROVIDER_AZURE_OPENAI,
         PROVIDER_COHERE,
         PROVIDER_GOOGLE_GEMINI,
-        # PROVIDER_HUGGING_FACE,
         PROVIDER_OLLAMA,
-        PROVIDER_TOGETHER_AI,
-        PROVIDER_AZURE_OPENAI,
         PROVIDER_OPENROUTER,
+        PROVIDER_SAMBANOVA,
+        PROVIDER_TOGETHER_AI,
     }
     PROVIDER_ENV_KEYS = {
-        PROVIDER_COHERE: "COHERE_API_KEY",
-        PROVIDER_GOOGLE_GEMINI: "GOOGLE_API_KEY",
-        PROVIDER_HUGGING_FACE: "HUGGINGFACEHUB_API_TOKEN",
-        PROVIDER_AZURE_OPENAI: "AZURE_OPENAI_API_KEY",
-        PROVIDER_OPENROUTER: "OPENROUTER_API_KEY",
-        PROVIDER_TOGETHER_AI: "TOGETHER_API_KEY",
+        PROVIDER_ANTHROPIC: 'ANTHROPIC_API_KEY',
+        PROVIDER_COHERE: 'COHERE_API_KEY',
+        PROVIDER_GOOGLE_GEMINI: 'GOOGLE_API_KEY',
+        PROVIDER_AZURE_OPENAI: 'AZURE_OPENAI_API_KEY',
+        PROVIDER_OPENROUTER: 'OPENROUTER_API_KEY',
+        PROVIDER_SAMBANOVA: 'SAMBANOVA_API_KEY',
+        PROVIDER_TOGETHER_AI: 'TOGETHER_API_KEY',
     }
     PROVIDER_REGEX = re.compile(r'\[(.*?)\]')
     VALID_MODELS = {
+        '[an]claude-haiku-4-5': {
+            'description': 'faster, detailed',
+            'max_new_tokens': 8192,
+            'paid': True,
+        },
         '[az]azure/open-ai': {
             'description': 'faster, detailed',
             'max_new_tokens': 8192,
@@ -76,16 +95,6 @@ class GlobalConfig:
             'max_new_tokens': 8192,
             'paid': True,
         },
-        # '[hf]mistralai/Mistral-7B-Instruct-v0.2': {
-        #     'description': 'faster, shorter',
-        #     'max_new_tokens': 8192,
-        #     'paid': False,
-        # },
-        # '[hf]mistralai/Mistral-Nemo-Instruct-2407': {
-        #     'description': 'longer response',
-        #     'max_new_tokens': 8192,
-        #     'paid': False,
-        # },
         '[or]google/gemini-2.0-flash-001': {
             'description': 'Google Gemini-2.0-flash-001 (via OpenRouter)',
             'max_new_tokens': 8192,
@@ -96,6 +105,11 @@ class GlobalConfig:
             'max_new_tokens': 4096,
             'paid': True,
         },
+        '[sn]Llama-3.3-Swallow-70B-Instruct-v0.4': {
+            'description': 'fast, shorter',
+            'max_new_tokens': 8192,
+            'paid': True,
+        },
         '[to]deepseek-ai/DeepSeek-V3': {
             'description': 'slower, medium',
             'max_new_tokens': 8192,

src/slidedeckai/helpers/llm_helper.py

@@ -27,8 +27,8 @@ except ImportError:
 
 LLM_PROVIDER_MODEL_REGEX = re.compile(r'\[(.*?)\](.*)')
 OLLAMA_MODEL_REGEX = re.compile(r'[a-zA-Z0-9._:-]+$')
-# 94 characters long, only containing alphanumeric characters, hyphens, and underscores
-API_KEY_REGEX = re.compile(r'^[a-zA-Z0-9_-]{6,94}$')
+# 128 characters long, only containing alphanumeric characters, hyphens, and underscores
+API_KEY_REGEX = re.compile(r'^[a-zA-Z0-9_-]{6,128}$')
 
 
 logger = logging.getLogger(__name__)
@@ -42,7 +42,6 @@ def get_provider_model(provider_model: str, use_ollama: bool) -> Tuple[str, str]
     :param use_ollama: Whether Ollama is used (i.e., running in offline mode).
     :return: The provider and the model name; empty strings in case no matching pattern found.
     """
-
     provider_model = provider_model.strip()
 
     if use_ollama:
@@ -99,7 +98,6 @@ def is_valid_llm_provider_model(
     :param azure_api_version: Azure OpenAI API version.
     :return: `True` if the settings "look" OK; `False` otherwise.
     """
-
     if not provider or not model or provider not in GlobalConfig.VALID_PROVIDERS:
         return False
 
@@ -132,16 +130,7 @@ def get_litellm_model_name(provider: str, model: str) -> Optional[str]:
     :param model: The model name.
     :return: LiteLLM-compatible model name, or None if provider is not supported.
     """
-    provider_prefix_map = {
-        GlobalConfig.PROVIDER_HUGGING_FACE: 'huggingface',
-        GlobalConfig.PROVIDER_GOOGLE_GEMINI: 'gemini',
-        GlobalConfig.PROVIDER_AZURE_OPENAI: 'azure',
-        GlobalConfig.PROVIDER_OPENROUTER: 'openrouter',
-        GlobalConfig.PROVIDER_COHERE: 'cohere',
-        GlobalConfig.PROVIDER_TOGETHER_AI: 'together_ai',
-        GlobalConfig.PROVIDER_OLLAMA: 'ollama',
-    }
-    prefix = provider_prefix_map.get(provider)
+    prefix = GlobalConfig.LITELLM_PROVIDER_MAPPING.get(provider)
     if prefix:
         return f'{prefix}/{model}'
     # LiteLLM always expects a prefix for model names; if not found, return None
@@ -171,7 +160,6 @@ def stream_litellm_completion(
     :param azure_api_version: Azure OpenAI API version.
     :return: Iterator of response chunks.
     """
-    
     if litellm is None:
         raise ImportError("LiteLLM is not installed. Please install it with: pip install litellm")
     
@@ -251,7 +239,6 @@ def get_litellm_llm(
     :param azure_api_version: Azure OpenAI API version.
     :return: A LiteLLM-compatible object for streaming; `None` in case of any error.
     """
-    
     if litellm is None:
         raise ImportError("LiteLLM is not installed. Please install it with: pip install litellm")
     

tests/unit/test_llm_helper.py

@@ -26,7 +26,6 @@ from slidedeckai.global_config import GlobalConfig
         ('invalid[]model', False, ('', '')),
         ('', False, ('', '')),
         ('[invalid]model', False, ('', '')),
-        ('[hf]mistral', False, ('', '')),  # hf is not in VALID_PROVIDERS
     ],
 )
 def test_get_provider_model(provider_model, use_ollama, expected):
@@ -107,7 +106,6 @@ def test_is_valid_llm_provider_model(
 @pytest.mark.parametrize(
     'provider, model, expected',
     [
-        (GlobalConfig.PROVIDER_HUGGING_FACE, 'mistral', 'huggingface/mistral'),
         (GlobalConfig.PROVIDER_GOOGLE_GEMINI, 'gemini-pro', 'gemini/gemini-pro'),
         (GlobalConfig.PROVIDER_OPENROUTER, 'openai/gpt-4', 'openrouter/openai/gpt-4'),
         (GlobalConfig.PROVIDER_COHERE, 'command', 'cohere/command'),
@@ -139,8 +137,8 @@ def test_stream_litellm_completion_success(mock_litellm):
     messages = [{'role': 'user', 'content': 'Say hello'}]
     result = list(
         stream_litellm_completion(
-            provider='hf',
-            model='mistral',
+            provider='gg',
+            model='gemini-2.5-flash-lite',
             messages=messages,
             max_tokens=100,
             api_key='test-key',
@@ -187,8 +185,8 @@ def test_stream_litellm_completion_error(mock_litellm):
     with pytest.raises(Exception) as exc_info:
         list(
             stream_litellm_completion(
-                provider='hf',
-                model='mistral',
+                provider='gg',
+                model='gemini-2.5-flash-lite',
                 messages=messages,
                 max_tokens=100,
                 api_key='test-key',
@@ -203,8 +201,8 @@ def test_get_litellm_llm(mock_stream):
     mock_stream.return_value = iter(['Hello', ' world'])
 
     llm = get_litellm_llm(
-        provider='hf',
-        model='mistral',
+        provider='gg',
+        model='gemini-2.5-flash-lite',
         max_new_tokens=100,
         api_key='test-key',
     )
@@ -254,8 +252,8 @@ def test_stream_litellm_completion_message_format(mock_litellm):
     messages = [{'role': 'user', 'content': 'Test'}]
     result = list(
         stream_litellm_completion(
-            provider='hf',
-            model='mistral',
+            provider='gg',
+            model='gemini-2.5-flash-lite',
             messages=messages,
             max_tokens=100,
             api_key='test-key',