adding deepseek

microsoft · Jan 6, 2025 · 4c72258 · 4c72258
1 parent 5c4b99f
commit 4c72258
Show file tree

Hide file tree

Showing 6 changed files with 41 additions and 5 deletions.
diff --git a/docs/src/content/docs/reference/cli/commands.md b/docs/src/content/docs/reference/cli/commands.md
@@ -16,7 +16,7 @@ Usage: genaiscript run [options] <script> [files...]
 Runs a GenAIScript against files.
 
 Options:
-  -p, --provider <string>                    Preferred LLM provider aliases (choices: "openai", "azure", "azure_serverless", "azure_serverless_models", "anthropic", "anthropic_bedrock", "google", "huggingface", "mistral", "alibaba", "github", "transformers", "ollama", "lmstudio", "jan", "llamafile", "litellm")
+  -p, --provider <string>                    Preferred LLM provider aliases (choices: "openai", "azure", "azure_serverless", "azure_serverless_models", "anthropic", "anthropic_bedrock", "google", "huggingface", "mistral", "alibaba", "github", "deepseek", "transformers", "ollama", "lmstudio", "jan", "llamafile", "litellm")
   -m, --model <string>                       'large' model alias (default)
   -sm, --small-model <string>                'small' alias model
   -vm, --vision-model <string>               'vision' alias model
@@ -92,8 +92,8 @@ Options:
                                       "azure_serverless_models", "anthropic",
                                       "anthropic_bedrock", "google",
                                       "huggingface", "mistral", "alibaba",
-                                      "github", "transformers", "ollama",
-                                      "lmstudio", "jan", "llamafile",
+                                      "github", "deepseek", "transformers",
+                                      "ollama", "lmstudio", "jan", "llamafile",
                                       "litellm")
   -m, --model <string>                'large' model alias (default)
   -sm, --small-model <string>         'small' alias model

diff --git a/packages/core/src/connection.ts b/packages/core/src/connection.ts
@@ -35,6 +35,8 @@ import {
     MODEL_PROVIDER_JAN,
     JAN_API_BASE,
     MODEL_PROVIDER_ANTHROPIC_BEDROCK,
+    MODEL_PROVIDER_DEEPSEEK,
+    DEEPSEEK_API_BASE,
 } from "./constants"
 import { host, runtimeHost } from "./host"
 import { parseModelIdentifier } from "./models"
@@ -412,6 +414,22 @@ export async function parseTokenFromEnv(
         } satisfies LanguageModelConfiguration
     }
 
+    if (provider === MODEL_PROVIDER_DEEPSEEK) {
+        const base =
+            findEnvVar(env, "DEEPSEEK", BASE_SUFFIX)?.value || DEEPSEEK_API_BASE
+        if (!URL.canParse(base)) throw new Error(`${base} must be a valid URL`)
+        const token = env.DEEPSEEK_API_KEY
+        if (!token) throw new Error("DEEPSEEK_API_KEY not configured")
+        return {
+            provider,
+            model,
+            base,
+            token,
+            type: "openai",
+            source: "env: DEEPSEEK_API_...",
+        }
+    }
+
     const prefixes = [
         tag ? `${provider}_${model}_${tag}` : undefined,
         provider ? `${provider}_${model}` : undefined,
@@ -493,7 +511,7 @@ export async function parseTokenFromEnv(
             provider,
             model,
             base,
-            token: "lmstudio",
+            token: "jan",
             type: "openai",
             source: "env: JAN_API_...",
         }

diff --git a/packages/core/src/constants.ts b/packages/core/src/constants.ts
@@ -122,6 +122,7 @@ export const GOOGLE_API_BASE =
 export const ALIBABA_BASE =
     "https://dashscope-intl.aliyuncs.com/compatible-mode/v1"
 export const MISTRAL_API_BASE = "https://api.mistral.ai/v1"
+export const DEEPSEEK_API_BASE = "https://api.deepseek.com/v1"
 
 export const PROMPTFOO_CACHE_PATH = ".genaiscript/cache/tests"
 export const PROMPTFOO_CONFIG_DIR = ".genaiscript/config/tests"
@@ -156,6 +157,7 @@ export const MODEL_PROVIDER_ALIBABA = "alibaba"
 export const MODEL_PROVIDER_MISTRAL = "mistral"
 export const MODEL_PROVIDER_LMSTUDIO = "lmstudio"
 export const MODEL_PROVIDER_JAN = "jan"
+export const MODEL_PROVIDER_DEEPSEEK = "deepseek"
 
 export const TRACE_FILE_PREVIEW_MAX_LENGTH = 240
 

diff --git a/packages/core/src/llms.json b/packages/core/src/llms.json
@@ -145,6 +145,17 @@
                 "reasoning_small": "o1-mini"
             }
         },
+        {
+            "id": "deepseek",
+            "detail": "DeepSeek Models",
+            "base": "https://api.deepseek.com/v1",
+            "bearerToken": true,
+            "aliases": {
+                "large": "deepseek-chat",
+                "small": "deepseek-chat",
+                "vision": "deepseek-chat"
+            }
+        },
         {
             "id": "transformers",
             "detail": "Hugging Face Transformers",
@@ -632,6 +643,11 @@
             "price_per_million_input_tokens": 0.8,
             "price_per_million_output_tokens": 4,
             "input_cache_token_rebate": 0.1
+        },
+        "deepseek:deepseek-chat": {
+            "price_per_million_input_tokens": 0.14,
+            "price_per_million_output_tokens": 0.28,
+            "input_cache_token_rebate": 0.1
         }
     }
 }
diff --git a/packages/core/src/lm.ts b/packages/core/src/lm.ts
@@ -7,7 +7,6 @@ import {
     MODEL_PROVIDER_ANTHROPIC,
     MODEL_PROVIDER_ANTHROPIC_BEDROCK,
     MODEL_PROVIDER_CLIENT,
-    MODEL_PROVIDER_JAN,
     MODEL_PROVIDER_OLLAMA,
     MODEL_PROVIDER_TRANSFORMERS,
     MODEL_PROVIDERS,

diff --git a/packages/core/src/types/prompt_template.d.ts b/packages/core/src/types/prompt_template.d.ts
@@ -175,6 +175,7 @@ type ModelType = OptionsOrString<
     | "alibaba:qwen-plus"
     | "alibaba:qwen2-72b-instruct"
     | "alibaba:qwen2-57b-a14b-instruct"
+    | "deepseek:deepseek-chat"
     | "transformers:onnx-community/Qwen2.5-0.5B-Instruct:q4"
     | "transformers:HuggingFaceTB/SmolLM2-1.7B-Instruct:q4f16"
 >