content-core 2.0.0__tar.gz → 2.0.2__tar.gz

{content_core-2.0.0 → content_core-2.0.2}/CHANGELOG.md

@@ -7,6 +7,22 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [2.0.2] - 2026-04-13
+
+### Fixed
+- LLM errors silently swallowed during summarization — errors now propagate with clear messages
+- Thinking models (e.g., Ollama Qwen 3.5) return empty content — now uses `cleaned_content` from Esperanto 2.20.1
+- Default LLM temperature changed from 0 to 0.5 and max_tokens from 600 to 4096 for compatibility with thinking models
+- LLM timeout increased to 120s for local model providers
+
+### Changed
+- Bumped esperanto dependency to >=2.20.1 (thinking model support, trailing slash fix)
+
+## [2.0.1] - 2026-04-13
+
+### Fixed
+- Summarize command fails via `uvx`/`pip install` — Jinja template was not packaged in the wheel
+
 ## [2.0.0] - 2026-04-11
 
 ### Added

{content_core-2.0.0 → content_core-2.0.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: content-core
-Version: 2.0.0
+Version: 2.0.2
 Summary: Extract what matters from any media source. Available as Python Library, macOS Service, CLI and MCP Server
 Author-email: LUIS NOVO <lfnovo@gmail.com>
 License-File: LICENSE
@@ -10,7 +10,7 @@ Requires-Dist: aiohttp>=3.13
 Requires-Dist: asciidoc>=10.2.1
 Requires-Dist: bs4>=0.0.2
 Requires-Dist: click>=8.3.0
-Requires-Dist: esperanto>=2.20.0
+Requires-Dist: esperanto>=2.20.1
 Requires-Dist: fast-ebook>=0.2.0
 Requires-Dist: fastmcp>=3.0.0
 Requires-Dist: firecrawl-py>=4.13.0

{content_core-2.0.0 → content_core-2.0.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "content-core"
-version = "2.0.0"
+version = "2.0.2"
 description = "Extract what matters from any media source. Available as Python Library, macOS Service, CLI and MCP Server"
 readme = "README.md"
 homepage = "https://github.com/lfnovo/content-core"
@@ -11,7 +11,7 @@ requires-python = ">=3.10"
 dependencies = [
     "aiohttp>=3.13",
     "bs4>=0.0.2",
-    "esperanto>=2.20.0",
+    "esperanto>=2.20.1",
     "jinja2>=3.1.6",
     "langdetect>=1.0.9",
     "loguru>=0.7.3",

{content_core-2.0.0 → content_core-2.0.2}/src/content_core/cli.py

@@ -69,7 +69,14 @@ def summarize(content, context):
     content = _get_content(content)
     # If content looks like a URL or file, extract first
     content = asyncio.run(_maybe_extract(content))
-    result = asyncio.run(summarize_fn(content, context))
+    try:
+        result = asyncio.run(summarize_fn(content, context))
+    except Exception as e:
+        click.echo(f"Error: {e}", err=True)
+        sys.exit(1)
+    if not result:
+        click.echo("Error: Summarization returned no result. Run with --debug for details.", err=True)
+        sys.exit(1)
     click.echo(result)
 
 

content_core-2.0.2/src/content_core/content/summary/core.py

@@ -0,0 +1,46 @@
+from functools import partial
+from typing import Optional
+
+from content_core.config import ContentCoreConfig
+from content_core.models import ModelFactory
+from content_core.templated_message import TemplatedMessageInput, templated_message
+
+SUMMARIZE_TEMPLATE = """\
+You are an AI assistant for a personal study platform.
+
+In this platform, your user collects various articles and content from the Internet for reference and study.
+
+Your role is to summarize the selected content as densely as possible, helping the reader extract maximum value from it without reading the full text.
+Focus solely on the content's value, avoiding unnecessary comments or messages.
+
+The summary should be dense, rich in characters, and designed to create a powerful vector representation.
+If the user provided additional context, follow its instructions. Otherwise, summary the whole content.
+
+Do not return any acknowledgments or greetings—only the summary.
+
+CONTENT:
+
+{{ content }}
+
+{% if context %}
+CONTEXT:
+
+User has provided the aditional context for your task:
+{{context}}
+{% endif %}
+
+SUMMARY:"""
+
+
+async def summarize(
+    content: str, context: str, config: Optional[ContentCoreConfig] = None
+) -> str:
+    model = ModelFactory.get_model("summary_model", config=config)
+    templated_message_fn = partial(templated_message, model=model)
+    response = await templated_message_fn(
+        TemplatedMessageInput(
+            user_prompt_text=SUMMARIZE_TEMPLATE,
+            data={"content": content, "context": context},
+        )
+    )
+    return response

{content_core-2.0.0 → content_core-2.0.2}/src/content_core/models.py

@@ -40,9 +40,9 @@ class ModelFactory:
         elif model_alias == "summary_model":
             provider = cfg.llm_provider
             model_name = cfg.summary_model or cfg.llm_model
-            return AIFactory.create_language(provider, model_name, config={})
+            return AIFactory.create_language(provider, model_name, config={"timeout": 120, "max_tokens": 4096})
         elif model_alias == "default_model":
-            return AIFactory.create_language(cfg.llm_provider, cfg.llm_model, config={})
+            return AIFactory.create_language(cfg.llm_provider, cfg.llm_model, config={"timeout": 120, "max_tokens": 4096})
         else:
             raise ValueError(f"Unknown model alias: {model_alias}")
 

{content_core-2.0.0 → content_core-2.0.2}/src/content_core/templated_message.py

@@ -5,7 +5,6 @@ from esperanto import LanguageModel
 from pydantic import BaseModel, Field
 
 from content_core.common.retry import retry_llm
-from content_core.logging import logger
 from content_core.models import ModelFactory
 
 
@@ -18,9 +17,9 @@ class TemplatedMessageInput(BaseModel):
     config: Dict = Field(
         description="The config for the LLM",
         default={
-            "temperature": 0,
+            "temperature": 0.5,
             "top_p": 1,
-            "max_tokens": 600,
+            "max_tokens": 4096,
         },
     )
 
@@ -29,7 +28,8 @@ class TemplatedMessageInput(BaseModel):
 async def _execute_llm_call(model: LanguageModel, msgs: list) -> str:
     """Internal function to execute LLM call - wrapped with retry logic."""
     result = await model.achat_complete(msgs)
-    return result.content
+    msg = result.choices[0].message
+    return getattr(msg, "cleaned_content", None) or msg.content
 
 
 async def templated_message(
@@ -63,8 +63,4 @@ async def templated_message(
         ).render(data=input.data)
         msgs.append({"role": "user", "content": user_prompt})
 
-    try:
-        return await _execute_llm_call(model, msgs)
-    except Exception as e:
-        logger.error(f"LLM call failed after retries: {e}")
-        return None
+    return await _execute_llm_call(model, msgs)

{content_core-2.0.0 → content_core-2.0.2}/uv.lock

@@ -775,7 +775,7 @@ wheels = [
 
 [[package]]
 name = "content-core"
-version = "2.0.0"
+version = "2.0.2"
 source = { editable = "." }
 dependencies = [
     { name = "ai-prompter" },
@@ -844,7 +844,7 @@ requires-dist = [
     { name = "click", specifier = ">=8.3.0" },
     { name = "crawl4ai", marker = "extra == 'crawl4ai'", specifier = ">=0.7.0" },
     { name = "docling", marker = "extra == 'docling'", specifier = ">=2.86.0" },
-    { name = "esperanto", specifier = ">=2.20.0" },
+    { name = "esperanto", specifier = ">=2.20.1" },
     { name = "fast-ebook", specifier = ">=0.2.0" },
     { name = "fastmcp", specifier = ">=3.0.0" },
     { name = "firecrawl-py", specifier = ">=4.13.0" },
@@ -1335,15 +1335,15 @@ wheels = [
 
 [[package]]
 name = "esperanto"
-version = "2.20.0"
+version = "2.20.1"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "httpx" },
     { name = "pydantic" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/bb/8c/eba93a2b7b1a4fc6812bcf0a83a69a3376fcaf5831fae04a0ad539baeff6/esperanto-2.20.0.tar.gz", hash = "sha256:53a16c539ed2f83e3400b4e56d5a20b7873df6cb926c1e6696a9611b3be928fb", size = 845084, upload-time = "2026-03-21T14:32:04.589Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/78/ae/7ff8cd45a2de81236722721417c06843abce3e2c3272edc2b6de1d800d08/esperanto-2.20.1.tar.gz", hash = "sha256:36707c5acdede3ea9c5de621f6dc4256d7e4b461b882cb63fff9e59b420e320c", size = 847075, upload-time = "2026-04-13T14:08:37.586Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/13/83/476dd4d7ee9f1b013ccd66e2518d150e66d3537736aed735b27230af0eb4/esperanto-2.20.0-py3-none-any.whl", hash = "sha256:767f95d08dd29e7d1950f4cc2ec908084e93ec61e945ce9446f001c54b86bb63", size = 203977, upload-time = "2026-03-21T14:32:05.676Z" },
+    { url = "https://files.pythonhosted.org/packages/93/ca/17e01c763e73c168868614746d1d694a10246ece5901b426355cd2c5dd83/esperanto-2.20.1-py3-none-any.whl", hash = "sha256:2cd258b7f1db0d68c06a61effabd7431d23a1a0040a9f413823d1542c521d7ea", size = 204252, upload-time = "2026-04-13T14:08:36.328Z" },
 ]
 
 [[package]]

content_core-2.0.0/src/content_core/content/summary/core.py

@@ -1,20 +0,0 @@
-from functools import partial
-from typing import Optional
-
-from content_core.config import ContentCoreConfig
-from content_core.models import ModelFactory
-from content_core.templated_message import TemplatedMessageInput, templated_message
-
-
-async def summarize(
-    content: str, context: str, config: Optional[ContentCoreConfig] = None
-) -> str:
-    model = ModelFactory.get_model("summary_model", config=config)
-    templated_message_fn = partial(templated_message, model=model)
-    response = await templated_message_fn(
-        TemplatedMessageInput(
-            user_prompt_template="content/summarize",
-            data={"content": content, "context": context},
-        )
-    )
-    return response