github/home-llm
1
1
Fork 0
mirror of https://github.com/acon96/home-llm.git synced 2026-01-06 20:33:54 -05:00
Code Issues Packages Projects Releases Wiki Activity

todo cleanup

Browse Source
This commit is contained in:
Alex O'Connell
2026-01-04 09:46:33 -05:00
parent 136a25f91b
commit 67de000326
2 changed files with 3 additions and 144 deletions
Download Patch File Download Diff File Expand all files Collapse all files

141
TEMPLATE_ERROR_FIX.md
View File

@@ -1,141 +0,0 @@
# Template Error Fix: LoggingUndefined Not JSON Serializable
## Error Message
```
Sorry, I had a problem with my template: TypeError: Type is not JSON serializable: LoggingUndefined
```
## Root Cause
The error occurs in the template rendering logic in `custom_components/llama_conversation/entity.py:662`. When your prompt template references a variable that doesn't exist, Jinja2 returns a special `LoggingUndefined` object. When Home Assistant tries to process this undefined value (during string formatting or JSON serialization), it fails because `LoggingUndefined` is not JSON serializable.
## The Problem Location
### In Your Template
The problematic line in your prompt template is:
```jinja2
Tools: {{ tools | to_json }}
```
### Why It Fails
The `tools` variable is **only available** when **"Enable Legacy Tool Calling"** is enabled in your configuration.
Looking at the code (`entity.py:645-655`), `tools` is only added to the template variables when `enable_legacy_tool_calling` is `True`:
```python
if enable_legacy_tool_calling:
if llm_api:
# ... tools are added here
render_variables["tools"] = tools
```
When this setting is disabled (which is the default), the template tries to access `{{ tools }}`, gets a `LoggingUndefined` object, and then `to_json` fails trying to serialize it.
## Available Template Variables
When rendering your prompt template, only these variables are available:
- `devices` - List of exposed devices
- `formatted_devices` - String representation of devices
- `response_examples` - ICL examples (if enabled)
- `tool_call_prefix` / `tool_call_suffix` - Tool call markers
- `tools` / `formatted_tools` - Available tools (⚠️ **only if legacy calling enabled**)
## Solutions
### Option 1: Enable Legacy Tool Calling (Quick Fix)
1. Go to your LLaMA Conversation integration settings
2. Find the "Enable Legacy Tool Calling" option
3. Enable it
4. Save and restart
### Option 2: Fix the Template (Recommended)
Replace the problematic line:
```jinja2
Tools: {{ tools | to_json }}
```
With one of these conditional versions:
#### Using `is defined` check:
```jinja2
{% if tools is defined %}Tools: {{ tools | to_json }}{% endif %}
```
#### Using `default()` filter:
```jinja2
Tools: {{ tools | default("No tools configured") | to_json }}
```
#### Using `default()` with empty list:
```jinja2
Tools: {{ tools | default([]) | to_json }}
```
## Complete Fixed Template
```jinja2
You are 'Al', a helpful AI Assistant that controls the devices in a house. Complete the following task as instructed with the information provided only.
The current time and date is {{ (as_timestamp(now()) | timestamp_custom("%I:%M %p on %A %B %d, %Y", True, "")) }}
{% if tools is defined %}Tools: {{ tools | to_json }}{% endif %}
Devices:
{% for device in devices | selectattr('area_id', 'none'): %}
{{ device.entity_id }} '{{ device.name }}' = {{ device.state }}{{ ([""] + device.attributes) | join(";") }}
{% endfor %}
{% for area in devices | rejectattr('area_id', 'none') | groupby('area_name') %}
## Area: {{ area.grouper }}
{% for device in area.list %}
{{ device.entity_id }} '{{ device.name }}' = {{ device.state }};{{ device.attributes | join(";") }}
{% endfor %}
{% endfor %}
{% for item in response_examples %}
{{ item.request }}
{{ item.response }}
<functioncall> {{ item.tool | to_json }}
{% endfor %}
```
## Prevention
To avoid this error in the future:
1. **Always check if optional variables are defined** before using them
2. **Use the `default()` filter** for variables that might not exist
3. **Stick to documented variables** listed in the integration documentation
4. **Test templates** after making changes
## Common Template Mistakes
❌ **Wrong:**
```jinja2
{{ some_typo }} # Undefined variable
{{ devises }} # Typo - should be "devices"
{{ device.missing_attr }} # Attribute doesn't exist
{{ tools | to_json }} # Not always available
```
✅ **Correct:**
```jinja2
{{ devices }} # Defined variable
{{ tools | default([]) }} # With fallback
{% if tools is defined %}{{ tools }}{% endif %} # With check
```
## Related Files
- `custom_components/llama_conversation/entity.py:662` - Template rendering
- `custom_components/llama_conversation/entity.py:637-660` - Available variables
- `custom_components/llama_conversation/entity.py:645-655` - Conditional `tools` variable
## Additional Notes
The `tools` variable is only populated when:
1. Legacy tool calling is enabled (`CONF_ENABLE_LEGACY_TOOL_CALLING = True`)
2. An LLM API is configured (`llm_api` is not `None`)
If you need tools in your template, ensure these conditions are met or use conditional checks as shown above.

6
TODO.md
View File

@@ -1,10 +1,10 @@
# TODO
- [ ] add examples of 'fixing' a failed tool call to the dataset
- [ ] add proper 'refusals' to the dataset (i.e. tool/device not available or device is already in the desired state)
- [x] add examples of 'fixing' a failed tool call to the dataset
- [x] add proper 'refusals' to the dataset (i.e. tool/device not available or device is already in the desired state)
- [ ] new model based on qwen3 0.6b, 1.7b and 4b
- [x] new model based on gemma3 270m
- [x] support AI task API
- [ ] vision support for remote backends
- [x] vision support for remote backends
- [ ] vision support for local backend (llama.cpp + llava)
- [ ] move llamacpp to a separate process because of all the crashing
- [ ] optional sampling parameters in options panel (don't pass to backend if not set)