next try

ai_integration/__manifest__.py

@@ -5,7 +5,7 @@
     'summary': 'Base module for AI integration',
     'description': """
 AI Integration Base
-==================
+===================
 This module provides the base framework for integrating various AI providers
 into Odoo. It includes:
 
@@ -18,6 +18,7 @@ into Odoo. It includes:
     'depends': [
         'base',
         'web',
+        'mail',
     ],
     'data': [
         'security/ai_security.xml',

ai_integration/models/__init__.py

@@ -1,4 +1,4 @@
-from .mixins import ai_mixin
+from .mixins.ai_base_mixin import AIBaseMixin
 from . import ai_generation_params
 from . import ai_model
 from . import ai_provider_interface

ai_integration/models/ai_generation_params.py

@@ -8,29 +8,35 @@ class AIGenerationParams(models.AbstractModel):
     temperature = fields.Float(
         string='Temperature',
         help='Controls randomness in generation. Higher values make output more random, lower values more deterministic.',
-        default=0.7)
+        default=0.7
+    )
     
     repeat_penalty = fields.Float(
         string='Repeat Penalty',
         help='Penalty for repeating tokens. Higher values make repetition less likely.',
-        default=1.1)
+        default=1.1
+    )
     
     max_tokens = fields.Integer(
         string='Max Tokens',
         help='Maximum number of tokens to generate.',
-        default=2048)
+        default=2048
+    )
     
     stop_sequences = fields.Char(
         string='Stop Sequences',
         help='Comma-separated list of sequences where generation should stop.',
-        default='')
+        default=''
+    )
     
     frequency_penalty = fields.Float(
         string='Frequency Penalty',
         help='Penalty for using frequent tokens. Higher values encourage using less frequent tokens.',
-        default=0.0)
+        default=0.0
+    )
     
     presence_penalty = fields.Float(
         string='Presence Penalty',
         help='Penalty for using tokens already in the text. Higher values encourage using new tokens.',
-        default=0.0)
+        default=0.0
+    )

ai_integration/models/ai_model.py

@@ -54,9 +54,9 @@ class AIModel(models.Model):
     )
     
     is_active = fields.Boolean(
-        string='Active',
+        string='Model Active',
         default=True,
-        help='Whether this model is active'
+        help='Whether this model is currently active and available for use'
     )
     
     context_window = fields.Integer(

ai_integration/models/ai_provider_instance.py

@@ -1,5 +1,6 @@
 # -*- coding: utf-8 -*-
 from odoo import models, fields, api, _
+from odoo.addons.mail.models.mail_thread import MailThread
 from odoo.exceptions import UserError
 
 
@@ -8,7 +9,7 @@ class AIProviderInstance(models.Model):
     _description = 'AI Provider Instance'
     _order = 'name'
     _check_company = False  # Disable automatic company checks
-    _inherit = ['ai.generation.params']
+    _inherit = ['mail.thread', 'ai.base.mixin']
 
     active = fields.Boolean(
         string='Active',

ai_integration/models/doc_models.md

@@ -0,0 +1,90 @@
+# Documentation des Modèles AI Integration
+
+## Vue d'ensemble
+
+Le module AI Integration fournit une infrastructure flexible pour intégrer différents fournisseurs d'IA dans Odoo. Il est conçu pour être extensible et permettre l'ajout facile de nouveaux fournisseurs.
+
+## Modèles Principaux
+
+### 1. AI Provider (`ai.provider`)
+- **Description**: Configuration de base des fournisseurs d'IA
+- **Champs principaux**:
+  - `name`: Nom du fournisseur
+  - `code`: Code technique unique
+  - `description`: Description détaillée
+  - `default_host`: Hôte par défaut
+  - `active`: État actif/inactif
+
+### 2. AI Provider Instance (`ai.provider.instance`)
+- **Description**: Instance spécifique d'un fournisseur d'IA
+- **Champs principaux**:
+  - `name`: Nom de l'instance
+  - `provider_id`: Fournisseur associé
+  - `provider_type`: Type de fournisseur
+  - `host`: Adresse de l'hôte
+  - `api_key`: Clé API (si nécessaire)
+  - `active`: État actif/inactif
+
+### 3. AI Model (`ai.model`)
+- **Description**: Modèles d'IA disponibles
+- **Champs principaux**:
+  - `name`: Nom du modèle
+  - `identifier`: Identifiant technique
+  - `provider_instance_id`: Instance du fournisseur
+  - `active`: État actif/inactif
+
+### 4. AI Model Stats (`ai.model.stats`)
+- **Description**: Statistiques d'utilisation des modèles
+- **Champs principaux**:
+  - `model_id`: Modèle associé
+  - `total_tokens`: Nombre total de tokens
+  - `total_requests`: Nombre total de requêtes
+  - `average_latency`: Latence moyenne
+
+## Mixin de Base
+
+### AI Base Mixin (`ai.base.mixin`)
+- **Description**: Mixin unifié pour l'intégration IA et les paramètres de génération
+- **Champs principaux**:
+  - `temperature`: Contrôle de la créativité (0.0 - 2.0)
+  - `top_p`: Sampling nucleus (0.0 - 1.0)
+  - `max_tokens`: Limite de tokens (1 - 32768)
+  - `stop_sequences`: Séquences d'arrêt
+  - `timeout`: Délai d'attente (1 - 300s)
+  - `retry_count`: Nombre de tentatives (0 - 5)
+  - `stream_response`: Activation du streaming
+- **Méthodes principales**:
+  - `_get_ai_provider_instance`: Obtenir l'instance du fournisseur
+  - `_get_ai_model`: Obtenir le modèle à utiliser
+  - `send_ai_message`: Envoyer un message à l'IA
+  - `_get_base_generation_params`: Obtenir les paramètres de génération
+
+## Configuration
+
+### 1. Res Config Settings
+- **Description**: Paramètres de configuration globaux
+- **Champs principaux**:
+  - `default_provider_instance_id`: Instance de fournisseur par défaut
+  - `default_model_id`: Modèle par défaut
+  - `ai_batch_size`: Taille du lot pour le traitement
+
+### 2. Res Company
+- **Description**: Extensions des paramètres de société
+- **Méthodes principales**:
+  - `_get_default_provider_instance`: Obtenir l'instance par défaut
+
+## Interfaces
+
+### AI Provider Interface (`ai.provider.interface`)
+- **Description**: Interface abstraite pour les fournisseurs d'IA
+- **Méthodes requises**:
+  - `send_message`: Envoyer un message
+  - `get_models`: Obtenir la liste des modèles
+  - `test_connection`: Tester la connexion
+
+## Notes d'Implémentation
+
+1. Tous les fournisseurs d'IA doivent implémenter `ai.provider.interface`
+2. Les instances de fournisseur héritent des paramètres de génération via `ai.generation.params`
+3. La configuration est hiérarchique : Global > Société > Instance
+4. Les statistiques sont collectées automatiquement pour chaque modèle

ai_integration/models/mixins/__init__.py

@@ -1,2 +1 @@
-from . import ai_mixin
+from . import ai_base_mixin
-from . import ai_generation_params

ai_integration/models/mixins/ai_base_mixin.py

@@ -0,0 +1,167 @@
+# -*- coding: utf-8 -*-
+from typing import List, Dict, Any, Optional
+from odoo import models, api, fields, _
+from odoo.exceptions import UserError
+import logging
+
+_logger = logging.getLogger(__name__)
+
+class AIBaseMixin(models.AbstractModel):
+    """Base mixin for AI integration providing both provider interaction and generation parameters.
+    
+    This mixin combines the functionality of message handling and generation parameters
+    into a single, cohesive interface for AI integration.
+    """
+    _name = 'ai.base.mixin'
+    _description = 'AI Integration Base Mixin'
+
+    # Basic Generation Parameters
+    temperature = fields.Float(
+        string='Temperature',
+        help='Sampling temperature. Range: [0.0 - 2.0]. Higher values make output more random, '
+             'lower values more deterministic.',
+        default=0.7,
+        digits=(3, 2))
+    
+    top_p = fields.Float(
+        string='Top P',
+        help='Nucleus sampling: limits cumulative probability of tokens to sample from. '
+             'Range: [0.0 - 1.0].',
+        default=0.9,
+        digits=(3, 2))
+    
+    max_tokens = fields.Integer(
+        string='Max Tokens',
+        help='Maximum number of tokens to generate. Range: [1 - 32768].',
+        default=2048)
+    
+    stop_sequences = fields.Char(
+        string='Stop Sequences',
+        help='Comma-separated list of sequences where the model should stop generating')
+
+    # System Settings
+    timeout = fields.Integer(
+        string='Timeout',
+        help='Request timeout in seconds. Range: [1 - 300].',
+        default=30)
+    
+    retry_count = fields.Integer(
+        string='Retry Count',
+        help='Number of times to retry failed requests. Range: [0 - 5].',
+        default=3)
+    
+    stream_response = fields.Boolean(
+        string='Stream Response',
+        help='Enable response streaming for real-time output.',
+        default=False)
+
+    def _get_base_generation_params(self):
+        """Get common generation parameters as a dictionary.
+        
+        Returns:
+            dict: Dictionary containing all generation parameters
+        """
+        self.ensure_one()
+        return {
+            'temperature': self.temperature,
+            'top_p': self.top_p,
+            'max_tokens': self.max_tokens,
+            'stop_sequences': self.stop_sequences.split(',') if self.stop_sequences else None,
+            'timeout': self.timeout,
+            'retry_count': self.retry_count,
+            'stream_response': self.stream_response,
+        }
+
+    def _get_ai_provider_instance(self, provider_instance_id=None):
+        """Get the AI provider instance to use.
+        
+        Args:
+            provider_instance_id: Optional specific provider instance to use
+            
+        Returns:
+            ai.provider.instance: The provider instance to use
+            
+        Raises:
+            UserError: If no provider instance is configured or available
+        """
+        if provider_instance_id:
+            instance = self.env['ai.provider.instance'].browse(provider_instance_id)
+            if not instance.exists():
+                raise UserError(_("Invalid provider instance"))
+        else:
+            provider_id = self.env['ir.config_parameter'].sudo().get_param(
+                'ai_integration.default_provider_instance_id')
+            if not provider_id:
+                raise UserError(_("No default AI provider instance configured"))
+            instance = self.env['ai.provider.instance'].browse(int(provider_id))
+            if not instance.exists():
+                raise UserError(_("Default provider instance not found"))
+        
+        if not instance.is_active:
+            raise UserError(_("The selected AI provider instance is not active"))
+            
+        return instance
+
+    def _get_ai_model(self, model_id=None, provider_instance=None):
+        """Get the AI model to use.
+        
+        Args:
+            model_id: Optional specific model to use
+            provider_instance: Optional provider instance (to avoid duplicate lookup)
+            
+        Returns:
+            ai.model: The model to use
+            
+        Raises:
+            UserError: If no model is configured or available
+        """
+        if not provider_instance:
+            provider_instance = self._get_ai_provider_instance()
+            
+        if model_id:
+            model = self.env['ai.model'].browse(model_id)
+            if not model.exists():
+                raise UserError(_("Invalid model"))
+            if model.provider_instance_id != provider_instance:
+                raise UserError(_("Model does not belong to the selected provider instance"))
+        else:
+            model_id = self.env['ir.config_parameter'].sudo().get_param(
+                'ai_integration.default_model_id')
+            if not model_id:
+                raise UserError(_("No default AI model configured"))
+            model = self.env['ai.model'].browse(int(model_id))
+            if not model.exists():
+                raise UserError(_("Default model not found"))
+                
+        if not model.is_active:
+            raise UserError(_("The selected AI model is not active"))
+            
+        return model
+
+    def send_ai_message(self, message: Dict[str, Any], provider_instance_id: Optional[int] = None,
+                       model_id: Optional[int] = None, **kwargs):
+        """Send a message to an AI provider instance.
+        
+        Args:
+            message: The message to send
+            provider_instance_id: Optional specific provider instance to use
+            model_id: Optional specific model to use
+            **kwargs: Additional provider-specific parameters
+            
+        Returns:
+            str: The response from the AI provider
+            
+        Raises:
+            UserError: If there's an error with the AI provider
+        """
+        provider_instance = self._get_ai_provider_instance(provider_instance_id)
+        model = self._get_ai_model(model_id, provider_instance)
+        
+        # Merge generation parameters with provider-specific parameters
+        params = {**self._get_base_generation_params(), **kwargs}
+        
+        try:
+            return provider_instance.send_message(message, model=model, **params)
+        except Exception as e:
+            _logger.error("Error sending message to AI provider: %s", str(e))
+            raise UserError(_("Failed to send message to AI provider: %s") % str(e))

ai_integration/models/mixins/ai_generation_params.py

@@ -1,64 +0,0 @@
-# -*- coding: utf-8 -*-
-from odoo import models, fields, api, _
-
-class AIGenerationParams(models.AbstractModel):
-    """Mixin for common AI generation parameters across different providers."""
-    _name = 'ai.generation.params'
-    _description = 'Common AI Generation Parameters'
-
-    # Basic Generation Parameters
-    temperature = fields.Float(
-        string='Temperature',
-        help='Sampling temperature. Range: [0.0 - 2.0]. Higher values make output more random, lower values more deterministic.',
-        default=0.7,
-        digits=(3, 2))
-    
-    top_p = fields.Float(
-        string='Top P',
-        help='Nucleus sampling: limits cumulative probability of tokens to sample from. Range: [0.0 - 1.0].',
-        default=0.9,
-        digits=(3, 2))
-    
-    max_tokens = fields.Integer(
-        string='Max Tokens',
-        help='Maximum number of tokens to generate. Range: [1 - 32768].',
-        default=2048)
-    
-    stop_sequences = fields.Char(
-        string='Stop Sequences',
-        help='Comma-separated list of sequences where the model should stop generating')
-
-    # System Settings
-    timeout = fields.Integer(
-        string='Timeout',
-        help='Request timeout in seconds. Range: [1 - 300].',
-        default=30)
-    
-    retry_count = fields.Integer(
-        string='Retry Count',
-        help='Number of times to retry failed requests. Range: [0 - 5].',
-        default=3)
-    
-    stream_response = fields.Boolean(
-        string='Stream Response',
-        help='Enable response streaming for real-time output.',
-        default=False)
-
-    def _get_base_generation_params(self):
-        """Get common generation parameters as a dictionary."""
-        self.ensure_one()
-        
-        params = {
-            'temperature': self.temperature,
-            'top_p': self.top_p,
-            'max_tokens': self.max_tokens,
-            'stream': self.stream_response,
-        }
-        
-        if self.stop_sequences:
-            params['stop'] = [
-                seq.strip() 
-                for seq in self.stop_sequences.split(',')
-            ]
-        
-        return params

ai_integration/models/mixins/ai_mixin.py

@@ -1,150 +0,0 @@
-# -*- coding: utf-8 -*-
-import logging
-from typing import List, Dict, Any, Optional
-from odoo import models, api, fields, _
-from odoo.exceptions import UserError
-
-_logger = logging.getLogger(__name__)
-
-
-class AIMixin(models.AbstractModel):
-    _name = 'ai.mixin'
-    _description = 'AI Integration Mixin'
-
-    def _get_ai_provider_instance(self, provider_instance_id=None):
-        """Get the AI provider instance to use.
-        
-        Args:
-            provider_instance_id: Optional specific provider instance to use
-            
-        Returns:
-            ai.provider.instance: The provider instance to use
-            
-        Raises:
-            UserError: If no provider instance is configured or available
-        """
-        if provider_instance_id:
-            instance = self.env['ai.provider.instance'].browse(provider_instance_id)
-            if not instance.exists():
-                raise UserError(_("Invalid provider instance"))
-        else:
-            provider_id = self.env['ir.config_parameter'].sudo().get_param('ai_integration.default_provider_instance_id')
-            if not provider_id:
-                raise UserError(_("No default AI provider instance configured"))
-            instance = self.env['ai.provider.instance'].browse(int(provider_id))
-            if not instance.exists():
-                raise UserError(_("Default provider instance not found"))
-        
-        if not instance.is_active:
-            raise UserError(_("The selected AI provider instance is not active"))
-            
-        return instance
-
-    def _get_ai_model(self, model_id=None, provider_instance=None):
-        """Get the AI model to use.
-        
-        Args:
-            model_id: Optional specific model to use
-            provider_instance: Optional provider instance (to avoid duplicate lookup)
-            
-        Returns:
-            ai.model: The model to use
-            
-        Raises:
-            UserError: If no model is configured or available
-        """
-        provider_instance = provider_instance or self._get_ai_provider_instance()
-        
-        if model_id:
-            model = self.env['ai.model'].browse(model_id)
-            if not model.exists():
-                raise UserError(_("Invalid AI model"))
-            if model.provider_instance_id != provider_instance:
-                raise UserError(_("The specified model does not belong to the selected provider instance"))
-        else:
-            model_id = self.env['ir.config_parameter'].sudo().get_param('ai_integration.default_model_id')
-            if not model_id:
-                raise UserError(_("No default AI model configured"))
-            model = self.env['ai.model'].browse(int(model_id))
-            if not model.exists():
-                raise UserError(_("Default AI model not found"))
-        
-        if not model.is_active:
-            raise UserError(_("The selected AI model is not active"))
-            
-        return model
-
-    def send_ai_message(self, message: Dict[str, Any], provider_instance_id: Optional[int] = None,
-                       model_id: Optional[int] = None, **kwargs) -> str:
-        """Send a message to an AI provider instance.
-        
-        Args:
-            message: The message to send
-            provider_instance_id: Optional specific provider instance to use
-            model_id: Optional specific model to use
-            **kwargs: Additional provider-specific parameters
-            
-        Returns:
-            str: The response from the AI provider
-            
-        Raises:
-            UserError: If there's an error with the AI provider
-        """
-        try:
-            instance = self._get_ai_provider_instance(provider_instance_id)
-            model = self._get_ai_model(model_id, instance)
-            return instance.send_message(message, model, **kwargs)
-        except Exception as e:
-            _logger.error("Error sending AI message: %s", str(e))
-            raise UserError(_("Error communicating with AI provider: %s", str(e)))
-
-    def process_batch_ai(self, items: List[Any], processor_func: callable,
-                        provider_instance_id: Optional[int] = None,
-                        model_id: Optional[int] = None, **kwargs) -> List[Any]:
-        """Process a batch of items using AI.
-        
-        Args:
-            items: List of items to process
-            processor_func: Function that processes each item and returns AI message
-            provider_instance_id: Optional specific provider instance to use
-            model_id: Optional specific model to use
-            **kwargs: Additional parameters passed to processor_func
-            
-        Returns:
-            List[Any]: List of processed results
-            
-        Example:
-            def _process_item(item, **kwargs):
-                return {'role': 'user', 'content': f'Analyze: {item.name}'}
-                
-            results = self.process_batch_ai(items, _process_item)
-        """
-        if not items:
-            return []
-
-        company = self.env.company
-        batch_size = company.ai_batch_size or 10
-        results = []
-
-        for i in range(0, len(items), batch_size):
-            batch = items[i:i + batch_size]
-            batch_messages = [processor_func(item, **kwargs) for item in batch]
-            
-            for message in batch_messages:
-                result = self.send_ai_message(
-                    message,
-                    provider_instance_id=provider_instance_id,
-                    model_id=model_id
-                )
-                results.append(result)
-
-        return results
-            
-    def _prepare_ai_message(self, **kwargs):
-        """Prepare a message to send to the AI provider.
-        This method should be implemented by models using this mixin.
-        
-        Returns:
-            dict: The prepared message
-        """
-        raise NotImplementedError(_("Method _prepare_ai_message must be implemented by models using AI mixin"))

ai_integration/security/ir.model.access.csv

@@ -7,3 +7,5 @@ access_ai_model_stats_user,ai.model.stats.user,model_ai_model_stats,base.group_u
 access_ai_model_stats_system,ai.model.stats.system,model_ai_model_stats,base.group_system,1,1,1,1
 access_ai_generation_params_user,ai.generation.params.user,model_ai_generation_params,base.group_user,1,0,0,0
 access_ai_generation_params_system,ai.generation.params.system,model_ai_generation_params,base.group_system,1,1,1,1
+access_ai_provider_user,ai.provider.user,model_ai_provider,base.group_user,1,0,0,0
+access_ai_provider_system,ai.provider.system,model_ai_provider,base.group_system,1,1,1,1

ollama_ai_integration/__manifest__.py

@@ -5,7 +5,7 @@
     'summary': 'Integration with Ollama AI models',
     'description': """
 Ollama Integration
-=================
+==================
 This module provides integration with Ollama, allowing you to use local AI models
 in your Odoo instance. Features include:
 
@@ -16,10 +16,12 @@ in your Odoo instance. Features include:
 """,
     'author': 'Bemade',
     'website': 'https://www.bemade.org',
-    'depends': ['ai_integration'],
+    'depends': [
+        'ai_integration'
+    ],
     'data': [
         'data/ollama_provider.xml',
-        'views/ollama_views.xml',
+        # 'views/ollama_views.xml',
         'views/ollama_stats_views.xml',
         'security/ir.model.access.csv',
     ],

ollama_ai_integration/models/__init__.py

@@ -1,4 +1,26 @@
+"""Ollama AI Integration Models Package.
+
+This package contains all the model definitions required for integrating
+Ollama AI with Odoo's AI framework. The models are loaded in a specific
+order to handle dependencies correctly.
+
+Module Structure:
+1. ollama_provider_mixin - Base configuration and parameter definitions
+2. ollama_provider - Core Ollama API integration implementation
+3. ollama_model_stats - Usage statistics and performance tracking
+4. ai_provider_instance - Instance-specific configuration and management
+
+Note: The import order is important to avoid circular dependencies.
+"""
+
+# Base Configuration
 from . import ollama_provider_mixin
-from . import ai_provider_ollama
+
-from . import ollama_instance
+# Core Implementation
+from . import ollama_provider
+
+# Statistics and Monitoring
 from . import ollama_model_stats
+
+# Instance Management
+from . import ai_provider_instance

ollama_ai_integration/models/ai_provider_instance.py

@@ -0,0 +1,131 @@
+from odoo import models, fields, api, _
+
+class AIProviderInstance(models.Model):
+    """Extends the AI Provider Instance model to support Ollama-specific configuration.
+    
+    This model inherits from both ai.provider.instance and ollama.provider.mixin to:
+    1. Add Ollama-specific fields (num_ctx, temperature, etc.)
+    2. Handle field visibility based on provider_type
+    3. Manage field cleanup when switching providers
+    
+    Note: This extends the base ai.provider.instance model instead of creating
+    a new one to ensure seamless integration with the core AI framework.
+    """
+    _name = 'ai.provider.instance'
+    _inherit = ['ollama.provider.mixin', 'mail.thread']
+    _description = 'AI Provider Instance'
+    
+    # Basic Fields
+    name = fields.Char(
+        string='Name',
+        required=True,
+        tracking=True,
+        help='Name of this AI provider instance')
+        
+    active = fields.Boolean(
+        string='Active',
+        default=True,
+        tracking=True,
+        help='Whether this provider instance is active')
+        
+    host = fields.Char(
+        string='Host',
+        required=True,
+        default='http://localhost:11434',
+        tracking=True,
+        help='Ollama server host URL')
+        
+    company_id = fields.Many2one(
+        'res.company',
+        string='Company',
+        required=True,
+        default=lambda self: self.env.company,
+        help='Company this provider instance belongs to')
+
+    @api.onchange('provider_type')
+    def _onchange_provider_type(self):
+        """Automatically clear Ollama-specific fields when switching provider type.
+        
+        This ensures that Ollama configuration is only kept when the provider
+        type is 'ollama'. When switching to another provider, all Ollama-specific
+        fields are reset to their default values to avoid confusion.
+        """
+        if self.provider_type != 'ollama':
+            self.update({
+                'num_ctx': False,  # Context length
+                'temperature': False,  # Sampling temperature
+                'top_p': False,  # Nucleus sampling threshold
+                'top_k': False,  # Top-k sampling threshold
+                'repeat_penalty': False,  # Penalty for repeated tokens
+            })
+
+    def test_connection(self):
+        """Test the connection to the Ollama server.
+        
+        This method attempts to connect to the Ollama server and verify
+        that it is responding correctly. It will raise a user-friendly
+        error if the connection fails.
+        
+        Returns:
+            dict: Action to display success message
+        """
+        self.ensure_one()
+        if self.provider_type != 'ollama':
+            return
+            
+        try:
+            # Try to list models as a basic connectivity test
+            self.env['ai.provider.ollama']._get_models(self)
+            return {
+                'type': 'ir.actions.client',
+                'tag': 'display_notification',
+                'params': {
+                    'title': _('Success'),
+                    'message': _('Successfully connected to Ollama server'),
+                    'sticky': False,
+                    'type': 'success',
+                }
+            }
+        except Exception as e:
+            raise UserError(_('Connection test failed: %s', str(e)))
+
+    def sync_models(self):
+        """Synchronize available models from the Ollama server.
+        
+        This method fetches the list of available models from the Ollama
+        server and creates or updates the corresponding AI model records
+        in Odoo.
+        
+        Returns:
+            dict: Action to display success message
+        """
+        self.ensure_one()
+        if self.provider_type != 'ollama':
+            return
+            
+        try:
+            provider = self.env['ai.provider.ollama']
+            models = provider._get_models(self)
+            
+            for model_data in models:
+                # Create or update AI model record
+                self.env['ai.model'].create_or_update({
+                    'name': model_data['name'],
+                    'identifier': model_data['id'],
+                    'provider_instance_id': self.id,
+                    'model_type': 'text',
+                    'active': True,
+                })
+                
+            return {
+                'type': 'ir.actions.client',
+                'tag': 'display_notification',
+                'params': {
+                    'title': _('Success'),
+                    'message': _('Successfully synchronized %d models', len(models)),
+                    'sticky': False,
+                    'type': 'success',
+                }
+            }
+        except Exception as e:
+            raise UserError(_('Model synchronization failed: %s', str(e)))

ollama_ai_integration/models/ai_provider_ollama.py.bak

@@ -1,6 +1,7 @@
 from odoo import models, fields, api, _
+from .ollama_provider_mixin import OllamaProviderMixin
 
-class OllamaProvider(models.Model):
+class OllamaAIProvider(models.Model, OllamaProviderMixin):
     _name = 'ai.provider.ollama'
     _description = 'Ollama AI Provider'
     _inherit = ['ai.provider.interface']

ollama_ai_integration/models/ollama_model_stats.py

@@ -2,10 +2,28 @@ from odoo import models, fields, api
 from datetime import datetime, timedelta
 
 class OllamaModelStats(models.Model):
+    """Tracks and stores daily usage statistics for Ollama AI models.
+    
+    This model maintains detailed daily statistics for each Ollama model,
+    including request counts, token usage, response times, and error rates.
+    It inherits from ai.model.stats for base statistics functionality.
+    
+    Key Features:
+    - Daily usage tracking per model
+    - Performance metrics collection
+    - Error rate monitoring
+    - Version tracking for model updates
+    
+    Technical Details:
+    - One stat entry per model per day (enforced by SQL constraint)
+    - Automatic version tracking from Ollama API
+    - Aggregated statistics calculation
+    - Ordered by date for easy historical analysis
+    """
     _name = 'ollama.model.stats'
     _description = 'Ollama Model Usage Statistics'
     _inherit = ['ai.model.stats']
-    _order = 'date desc'
+    _order = 'date desc'  # Most recent stats first
 
     model_id = fields.Many2one('ai.model', string='Model', required=True, ondelete='cascade')
     date = fields.Date(string='Date', required=True, default=fields.Date.context_today)
@@ -20,7 +38,23 @@ class OllamaModelStats(models.Model):
     ]
 
     def _update_stats(self, model, tokens, response_time, error=False):
-        """Update statistics for a model."""
+        """Update daily statistics for a specific model.
+        
+        This method handles the creation or update of daily statistics entries.
+        It maintains running averages and cumulative counts for various metrics.
+        
+        Args:
+            model (ai.model): The model record being tracked
+            tokens (int): Number of tokens in the current request
+            response_time (float): Response time in milliseconds
+            error (bool): Whether this request resulted in an error
+            
+        Technical Notes:
+            - Creates new stat entry if none exists for today
+            - Updates running averages for response time
+            - Fetches and stores model version from Ollama API
+            - Maintains cumulative counts for requests and errors
+        """
         today = fields.Date.context_today(self)
         stats = self.search([
             ('model_id', '=', model.id),

ollama_ai_integration/models/ollama_provider.py

@@ -1,15 +1,70 @@
 import json
 import logging
 import requests
-from odoo import models, fields, _
+from odoo import models, fields, api, _
 from odoo.exceptions import UserError
 
 _logger = logging.getLogger(__name__)
 
 class OllamaProvider(models.Model):
+    """Main Ollama AI Provider implementation.
+    
+    This model implements the core functionality for interacting with Ollama's API,
+    including model management, text generation, and error handling.
+    
+    Key Responsibilities:
+    - Model discovery and validation
+    - API communication and error handling
+    - Request formatting and response parsing
+    - Resource management and cleanup
+    
+    Technical Details:
+    - Implements the ai.provider.interface for standardized AI provider integration
+    - Uses Ollama's HTTP API for all operations
+    - Handles both synchronous and asynchronous requests
+    - Provides detailed error messages for troubleshooting
+    """
     _name = 'ai.provider.ollama'
     _description = 'Ollama AI Provider'
     _inherit = ['ai.provider.interface']
+    
+    @api.model
+    def _get_models(self, instance):
+        """Get list of available models from Ollama server.
+        
+        Args:
+            instance (ai.provider.instance): Provider instance to get models for
+            
+        Returns:
+            list: List of model dictionaries with keys:
+                - name: Model name
+                - id: Model identifier
+                - details: Additional model metadata
+                
+        Raises:
+            UserError: If unable to connect or retrieve models
+        """
+        try:
+            response = requests.get(f"{instance.host}/api/tags")
+            response.raise_for_status()
+            
+            models_data = response.json().get('models', [])
+            return [{
+                'name': model['name'],
+                'id': model['name'],
+                'details': model
+            } for model in models_data]
+            
+        except requests.exceptions.RequestException as e:
+            raise UserError(_('Failed to connect to Ollama server: %s', str(e)))
+        except (KeyError, ValueError) as e:
+            raise UserError(_('Invalid response from Ollama server: %s', str(e)))
+
+    # API Configuration
+    timeout = fields.Integer(
+        string='Timeout',
+        default=30,
+        help='API request timeout in seconds')
 
     def _get_provider_type(self):
         return 'ollama'

ollama_ai_integration/models/ollama_provider_mixin.py

@@ -1,25 +1,96 @@
 from odoo import models, fields, api, _
 
 class OllamaProviderMixin(models.AbstractModel):
+    """Mixin model that provides Ollama-specific configuration parameters.
+    
+    This mixin is designed to be inherited by models that need to interact with
+    the Ollama AI provider. It provides all the necessary fields and methods
+    for configuring and interacting with Ollama's API.
+    
+    Key Features:
+    - Provider type selection and validation
+    - Context window configuration
+    - Advanced sampling parameters (temperature, top-k, top-p)
+    - Token generation controls
+    
+    Technical Details:
+    - Inherits from ai.generation.params for base AI generation parameters
+    - Implements Ollama-specific API parameters
+    - Provides default values optimized for general use cases
+    """
     _name = 'ollama.provider.mixin'
     _description = 'Ollama Provider Configuration Mixin'
     _inherit = ['ai.generation.params']
 
+    # Provider Configuration
     provider_type = fields.Selection(
-        selection_add=[('ollama', 'Ollama')],
+        selection=[('ollama', 'Ollama')],
-        ondelete={'ollama': 'cascade'})
+        string='Provider Type',
+        required=True,
+        default='ollama',
+        help='Type of AI provider - Must be Ollama for this configuration')
     
-    # Ollama-specific Parameters
+    # Model Parameters
+    model_name = fields.Char(
+        string='Model Name',
+        help='Name of the Ollama model to use (e.g. llama2, mistral, codellama)',
+        required=True,
+        default='llama2')
+        
+    # Context Window Configuration
     num_ctx = fields.Integer(
         string='Context Length',
-        help='Maximum number of tokens to consider for context. Range: [0 - 32768].',
+        help='Maximum number of tokens to consider for context. A larger context window allows '
+             'the model to access more historical information but requires more memory. '
+             'Range: [0 - 32768].',
         default=4096)
-    
+        
-    # Advanced Sampling Parameters
+    # Generation Parameters
+    temperature = fields.Float(
+        string='Temperature',
+        help='Controls randomness in the output. Higher values make the output more random, '
+             'while lower values make it more focused and deterministic. '
+             'Range: [0.0 - 2.0]',
+        default=0.8)
+        
+    top_p = fields.Float(
+        string='Top P',
+        help='Nucleus sampling: only consider the tokens whose cumulative probability exceeds '
+             'this value. Lower values make the output more focused. '
+             'Range: [0.0 - 1.0]',
+        default=0.9)
+        
     top_k = fields.Integer(
         string='Top K',
-        help='Limits the number of tokens to sample from. Range: [1 - 100].',
+        help='Only consider the top K tokens for text generation. Lower values make the '
+             'output more focused. Set to 0 to disable. '
+             'Range: [0 - 100]',
         default=40)
+        
+    repeat_penalty = fields.Float(
+        string='Repeat Penalty',
+        help='Penalty for repeating tokens. Higher values make the output less repetitive. '
+             'Range: [0.0 - 2.0]',
+        default=1.1)
+    
+    # Advanced Configuration
+    stop_sequences = fields.Char(
+        string='Stop Sequences',
+        help='Comma-separated list of sequences where the model should stop generating further tokens.')
+    
+    top_k = fields.Integer(
+        string='Top K',
+        help='Limits the cumulative probability of tokens to sample from. Only the top K '
+             'most likely tokens are considered for sampling at each step. '
+             'Range: [1 - 100].',
+        default=40)
+    
+    top_p = fields.Float(
+        string='Top P (Nucleus Sampling)',
+        help='Limits the cumulative probability of tokens to sample from. Only the most likely '
+             'tokens with total probability mass of top_p are considered. '
+             'Range: [0.0 - 1.0].',
+        default=0.9)
     
     min_p = fields.Float(
         string='Min P',

ollama_ai_integration/security/ir.model.access.csv

@@ -1,3 +1,5 @@
 id,name,model_id:id,group_id:id,perm_read,perm_write,perm_create,perm_unlink
 access_ollama_model_stats_user,ollama.model.stats.user,model_ollama_model_stats,base.group_user,1,0,0,0
 access_ollama_model_stats_manager,ollama.model.stats.manager,model_ollama_model_stats,base.group_system,1,1,1,1
+access_ai_provider_ollama_user,ai.provider.ollama.user,model_ai_provider_ollama,base.group_user,1,0,0,0
+access_ai_provider_ollama_manager,ai.provider.ollama.manager,model_ai_provider_ollama,base.group_system,1,1,1,1

ollama_ai_integration/views/ollama_views.xml

@@ -15,6 +15,9 @@
                     </button>
                 </div>
             </xpath>
+            <xpath expr="//field[@name='api_key']" position="replace">
+                <field name="api_key" invisible="1"/>
+            </xpath>
             <xpath expr="//notebook" position="inside">
                 <page string="Ollama Settings" name="ollama_settings"
                       invisible="provider_type != 'ollama'">