letta-nightly 0.11.0.dev20250808055434__py3-none-any.whl → 0.11.0.dev20250808104456__py3-none-any.whl

letta/constants.py

@@ -10,6 +10,7 @@ DEFAULT_TIMEZONE = "UTC"
 
 ADMIN_PREFIX = "/v1/admin"
 API_PREFIX = "/v1"
+OLLAMA_API_PREFIX = "/v1"
 OPENAI_API_PREFIX = "/openai"
 
 COMPOSIO_ENTITY_ENV_VAR_KEY = "COMPOSIO_ENTITY"
@@ -50,8 +51,9 @@ TOOL_CALL_ID_MAX_LEN = 29
 # Max steps for agent loop
 DEFAULT_MAX_STEPS = 50
 
-# minimum context window size
+# context window size
 MIN_CONTEXT_WINDOW = 4096
+DEFAULT_CONTEXT_WINDOW = 32000
 
 # number of concurrent embedding requests to sent
 EMBEDDING_BATCH_SIZE = 200
@@ -63,6 +65,7 @@ DEFAULT_MIN_MESSAGE_BUFFER_LENGTH = 15
 # embeddings
 MAX_EMBEDDING_DIM = 4096  # maximum supported embeding size - do NOT change or else DBs will need to be reset
 DEFAULT_EMBEDDING_CHUNK_SIZE = 300
+DEFAULT_EMBEDDING_DIM = 1024
 
 # tokenizers
 EMBEDDING_TO_TOKENIZER_MAP = {

letta/embeddings.py

@@ -139,10 +139,11 @@ class AzureOpenAIEmbedding:
 
 class OllamaEmbeddings:
 
+    # Uses OpenAI API standard
     # Format:
-    # curl http://localhost:11434/api/embeddings -d '{
+    # curl http://localhost:11434/v1/embeddings -d '{
     #   "model": "mxbai-embed-large",
-    #   "prompt": "Llamas are members of the camelid family"
+    #   "input": "Llamas are members of the camelid family"
     # }'
 
     def __init__(self, model: str, base_url: str, ollama_additional_kwargs: dict):
@@ -154,18 +155,18 @@ class OllamaEmbeddings:
         import httpx
 
         headers = {"Content-Type": "application/json"}
-        json_data = {"model": self.model, "prompt": text}
+        json_data = {"model": self.model, "input": text}
         json_data.update(self.ollama_additional_kwargs)
 
         with httpx.Client() as client:
             response = client.post(
-                f"{self.base_url}/api/embeddings",
+                f"{self.base_url}/embeddings",
                 headers=headers,
                 json=json_data,
             )
 
         response_json = response.json()
-        return response_json["embedding"]
+        return response_json["data"][0]["embedding"]
 
 
 class GoogleEmbeddings:

letta/schemas/providers/ollama.py

@@ -3,7 +3,7 @@ from typing import Literal
 import aiohttp
 from pydantic import Field
 
-from letta.constants import DEFAULT_EMBEDDING_CHUNK_SIZE
+from letta.constants import DEFAULT_EMBEDDING_CHUNK_SIZE, DEFAULT_CONTEXT_WINDOW, DEFAULT_EMBEDDING_DIM, OLLAMA_API_PREFIX
 from letta.log import get_logger
 from letta.schemas.embedding_config import EmbeddingConfig
 from letta.schemas.enums import ProviderCategory, ProviderType
@@ -12,8 +12,6 @@ from letta.schemas.providers.openai import OpenAIProvider
 
 logger = get_logger(__name__)
 
-ollama_prefix = "/v1"
-
 
 class OllamaProvider(OpenAIProvider):
     """Ollama provider that uses the native /api/generate endpoint
@@ -41,19 +39,30 @@ class OllamaProvider(OpenAIProvider):
                 response_json = await response.json()
 
         configs = []
-        for model in response_json["models"]:
-            context_window = await self._get_model_context_window(model["name"])
+        for model in response_json.get("models", []):
+            model_name = model["name"]
+            model_details = await self._get_model_details_async(model_name)
+            if not model_details or "completion" not in model_details.get("capabilities", []):
+                continue
+
+            context_window = None
+            model_info = model_details.get("model_info", {})
+            if architecture := model_info.get("general.architecture"):
+                if context_length := model_info.get(f"{architecture}.context_length"):
+                    context_window = int(context_length)
+
             if context_window is None:
-                print(f"Ollama model {model['name']} has no context window, using default 32000")
-                context_window = 32000
+                logger.warning(f"Ollama model {model_name} has no context window, using default {DEFAULT_CONTEXT_WINDOW}")
+                context_window = DEFAULT_CONTEXT_WINDOW
+
             configs.append(
                 LLMConfig(
-                    model=model["name"],
+                    model=model_name,
                     model_endpoint_type=ProviderType.ollama,
-                    model_endpoint=f"{self.base_url}{ollama_prefix}",
+                    model_endpoint=f"{self.base_url}{OLLAMA_API_PREFIX}",
                     model_wrapper=self.default_prompt_formatter,
                     context_window=context_window,
-                    handle=self.get_handle(model["name"]),
+                    handle=self.get_handle(model_name),
                     provider_name=self.name,
                     provider_category=self.provider_category,
                 )
@@ -73,25 +82,36 @@ class OllamaProvider(OpenAIProvider):
                 response_json = await response.json()
 
         configs = []
-        for model in response_json["models"]:
-            embedding_dim = await self._get_model_embedding_dim(model["name"])
+        for model in response_json.get("models", []):
+            model_name = model["name"]
+            model_details = await self._get_model_details_async(model_name)
+            if not model_details or "embedding" not in model_details.get("capabilities", []):
+                continue
+
+            embedding_dim = None
+            model_info = model_details.get("model_info", {})
+            if architecture := model_info.get("general.architecture"):
+                if embedding_length := model_info.get(f"{architecture}.embedding_length"):
+                    embedding_dim = int(embedding_length)
+
             if not embedding_dim:
-                print(f"Ollama model {model['name']} has no embedding dimension, using default 1024")
-                # continue
-                embedding_dim = 1024
+                logger.warning(f"Ollama model {model_name} has no embedding dimension, using default {DEFAULT_EMBEDDING_DIM}")
+                embedding_dim = DEFAULT_EMBEDDING_DIM
+
             configs.append(
                 EmbeddingConfig(
-                    embedding_model=model["name"],
+                    embedding_model=model_name,
                     embedding_endpoint_type=ProviderType.ollama,
-                    embedding_endpoint=f"{self.base_url}{ollama_prefix}",
+                    embedding_endpoint=f"{self.base_url}{OLLAMA_API_PREFIX}",
                     embedding_dim=embedding_dim,
                     embedding_chunk_size=DEFAULT_EMBEDDING_CHUNK_SIZE,
-                    handle=self.get_handle(model["name"], is_embedding=True),
+                    handle=self.get_handle(model_name, is_embedding=True),
                 )
             )
         return configs
 
-    async def _get_model_context_window(self, model_name: str) -> int | None:
+    async def _get_model_details_async(self, model_name: str) -> dict | None:
+        """Get detailed information for a specific model from /api/show."""
         endpoint = f"{self.base_url}/api/show"
         payload = {"name": model_name}
 
@@ -102,39 +122,7 @@ class OllamaProvider(OpenAIProvider):
                         error_text = await response.text()
                         logger.warning(f"Failed to get model info for {model_name}: {response.status} - {error_text}")
                         return None
-
-                    response_json = await response.json()
-                    model_info = response_json.get("model_info", {})
-
-                    if architecture := model_info.get("general.architecture"):
-                        if context_length := model_info.get(f"{architecture}.context_length"):
-                            return int(context_length)
-
+                    return await response.json()
         except Exception as e:
-            logger.warning(f"Failed to get model context window for {model_name} with error: {e}")
-
-        return None
-
-    async def _get_model_embedding_dim(self, model_name: str) -> int | None:
-        endpoint = f"{self.base_url}/api/show"
-        payload = {"name": model_name}
-
-        try:
-            async with aiohttp.ClientSession() as session:
-                async with session.post(endpoint, json=payload) as response:
-                    if response.status != 200:
-                        error_text = await response.text()
-                        logger.warning(f"Failed to get model info for {model_name}: {response.status} - {error_text}")
-                        return None
-
-                    response_json = await response.json()
-                    model_info = response_json.get("model_info", {})
-
-                    if architecture := model_info.get("general.architecture"):
-                        if embedding_length := model_info.get(f"{architecture}.embedding_length"):
-                            return int(embedding_length)
-
-        except Exception as e:
-            logger.warning(f"Failed to get model embedding dimension for {model_name} with error: {e}")
-
-        return None
+            logger.warning(f"Failed to get model details for {model_name} with error: {e}")
+            return None

{letta_nightly-0.11.0.dev20250808055434.dist-info → letta_nightly-0.11.0.dev20250808104456.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: letta-nightly
-Version: 0.11.0.dev20250808055434
+Version: 0.11.0.dev20250808104456
 Summary: Create LLM agents with long-term memory and custom tools
 License: Apache License
 Author: Letta Team

{letta_nightly-0.11.0.dev20250808055434.dist-info → letta_nightly-0.11.0.dev20250808104456.dist-info}/RECORD

@@ -16,12 +16,12 @@ letta/client/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 letta/client/streaming.py,sha256=UsDS_tDTsA3HgYryIDvGGmx_dWfnfQwtmEwLi4Z89Ik,4701
 letta/client/utils.py,sha256=VCGV-op5ZSmurd4yw7Vhf93XDQ0BkyBT8qsuV7EqfiU,2859
 letta/config.py,sha256=JFGY4TWW0Wm5fTbZamOwWqk5G8Nn-TXyhgByGoAqy2c,12375
-letta/constants.py,sha256=CIo3xpLsYr1EqYZJdrlDqPOoMmomtOlFYX4eV8uF-pk,15326
+letta/constants.py,sha256=8bOs4Ya6ZQO3kCJFoc9rFJX-NW-FlMGduXq8DdkpZeM,15404
 letta/data_sources/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 letta/data_sources/connectors.py,sha256=V8mUgE3V6CX-CcOyvkPSQ_ZWP2VtuqgTEXkCN1j0p68,7920
 letta/data_sources/connectors_helper.py,sha256=oQpVlc-BjSz9sTZ7sp4PsJSXJbBKpZPi3Dam03CURTQ,3376
 letta/data_sources/redis_client.py,sha256=Lz9hjJL9S7yd_qsvbuwqrGL3GwZe-qOc0uZm9FqNm4M,10688
-letta/embeddings.py,sha256=5bxVQOJHSXezbHH-htZDfzce_ZTEn_WY5E07PS09HO8,10555
+letta/embeddings.py,sha256=WNSUAMEWLMDcQg_2iLap8ApmOnCT98JytnBHWuv_rCg,10590
 letta/errors.py,sha256=DiskTVZiSZ4jbjEBYw7TDvuslnYhcXSXX2LUzmMbydo,7632
 letta/functions/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 letta/functions/ast_parsers.py,sha256=0dXAN4qx3pWL_Y0aoEkaBpMKwI-kpoLEJftjW3v2I4E,5031
@@ -290,7 +290,7 @@ letta/schemas/providers/groq.py,sha256=AquJQH-Y5-s75Nj2_X7xavuWUu5F2bSvHjAZ1Gfpe
 letta/schemas/providers/letta.py,sha256=50VcmTMm8OEQjZgdlLypa4QmNPe0mJycpZfbxxSb_ts,1611
 letta/schemas/providers/lmstudio.py,sha256=Hi8Nir96B5FCQ8cVT-mWl7nLX0Z2-RIxUANAqRGG-zo,4323
 letta/schemas/providers/mistral.py,sha256=EjFF6YcfN5jBjCfnZw3ECv_3qYuG0HVb7B0VoYk-jKU,1866
-letta/schemas/providers/ollama.py,sha256=iVx9xxrulG9ohbhk4kMtfryCnMgmynWg_NL4SesGX5U,6253
+letta/schemas/providers/ollama.py,sha256=6uMKhK7lq_f9-k0oFNO7NDhBdpv57HxyjWlYTuuYYsE,5970
 letta/schemas/providers/openai.py,sha256=jlnMu3t1_IHWT4dGn8zZlbITl6wQl2X9onn_B2ZhV48,11051
 letta/schemas/providers/together.py,sha256=2zFca6Jy08r1ANrdvtlSIduyDr8ek9Tt1yYiz1S-5g8,3422
 letta/schemas/providers/vllm.py,sha256=CwM260cxWLkviVzY4wwkw4NmDAK69fv531AofRGa9JA,2480
@@ -456,8 +456,8 @@ letta/templates/summary_request_text.j2,sha256=ZttQwXonW2lk4pJLYzLK0pmo4EO4EtUUI
 letta/templates/template_helper.py,sha256=HkG3zwRc5NVGmSTQu5PUTpz7LevK43bzXVaQuN8urf0,1634
 letta/types/__init__.py,sha256=hokKjCVFGEfR7SLMrtZsRsBfsC7yTIbgKPLdGg4K1eY,147
 letta/utils.py,sha256=Fwwe2imHRamc_kucAATo8NXhwDG5NBoOIYmBaERXUhM,38384
-letta_nightly-0.11.0.dev20250808055434.dist-info/LICENSE,sha256=mExtuZ_GYJgDEI38GWdiEYZizZS4KkVt2SF1g_GPNhI,10759
-letta_nightly-0.11.0.dev20250808055434.dist-info/METADATA,sha256=oGIMFfhX6qRD55DEP_UdK596w0AMrKXiUlo-VWTAJfw,23281
-letta_nightly-0.11.0.dev20250808055434.dist-info/WHEEL,sha256=FMvqSimYX_P7y0a7UY-_Mc83r5zkBZsCYPm7Lr0Bsq4,88
-letta_nightly-0.11.0.dev20250808055434.dist-info/entry_points.txt,sha256=2zdiyGNEZGV5oYBuS-y2nAAgjDgcC9yM_mHJBFSRt5U,40
-letta_nightly-0.11.0.dev20250808055434.dist-info/RECORD,,
+letta_nightly-0.11.0.dev20250808104456.dist-info/LICENSE,sha256=mExtuZ_GYJgDEI38GWdiEYZizZS4KkVt2SF1g_GPNhI,10759
+letta_nightly-0.11.0.dev20250808104456.dist-info/METADATA,sha256=Z0sSJdcPjhEtwrK7ZZNPUc_fJm1Yyk3mf7-uMTHylHQ,23281
+letta_nightly-0.11.0.dev20250808104456.dist-info/WHEEL,sha256=FMvqSimYX_P7y0a7UY-_Mc83r5zkBZsCYPm7Lr0Bsq4,88
+letta_nightly-0.11.0.dev20250808104456.dist-info/entry_points.txt,sha256=2zdiyGNEZGV5oYBuS-y2nAAgjDgcC9yM_mHJBFSRt5U,40
+letta_nightly-0.11.0.dev20250808104456.dist-info/RECORD,,