Odoo-Modules/fusion_accounting_bank_rec/models/fusion_reconcile_suggestion.py

"""Persisted AI suggestions for bank line reconciliations.

One row per (statement_line, candidate_match). The OWL widget reads these
to render confidence badges; users accept/reject which feeds back into
the pattern learning system.

The AI never writes account.partial.reconcile directly — it writes
suggestions here, and the user (or batch-accept action) approves them
through the engine's accept_suggestion() method.
"""

import logging

from odoo import api, fields, models

_logger = logging.getLogger(__name__)


class FusionReconcileSuggestion(models.Model):
    _name = "fusion.reconcile.suggestion"
    _description = "AI-generated bank reconciliation suggestion"
    _order = "statement_line_id, confidence desc"

    company_id = fields.Many2one('res.company', required=True, index=True,
                                   default=lambda self: self.env.company)
    statement_line_id = fields.Many2one('account.bank.statement.line',
                                          required=True, index=True, ondelete='cascade')

    # The proposal
    proposed_move_line_ids = fields.Many2many('account.move.line',
                                                string="Proposed matches")
    proposed_write_off_amount = fields.Monetary(currency_field='currency_id')
    proposed_write_off_account_id = fields.Many2one('account.account')
    currency_id = fields.Many2one('res.currency',
                                    related='statement_line_id.currency_id',
                                    store=True)

    # Scoring
    confidence = fields.Float(required=True)
    confidence_band = fields.Selection([
        ('high', 'High (>=95%)'),
        ('medium', 'Medium (70-94%)'),
        ('low', 'Low (50-69%)'),
        ('none', 'No confidence (<50%)'),
    ], compute='_compute_band', store=True)
    rank = fields.Integer(help="1 = top suggestion, 2-N = alternatives")
    reasoning = fields.Text(help="Human-readable explanation")

    # Feature breakdown (for transparency + future learning)
    score_amount_match = fields.Float()
    score_partner_pattern = fields.Float()
    score_precedent_similarity = fields.Float()
    score_ai_rerank = fields.Float()

    # Provenance
    generated_at = fields.Datetime(default=fields.Datetime.now)
    generated_by = fields.Selection([
        ('cron_batch', 'Batch cron'),
        ('on_demand', 'User refreshed alternatives'),
        ('on_open', 'Widget opened (lazy)'),
    ])
    provider_used = fields.Char(
        help="e.g. 'claude_sonnet_4_5', 'lmstudio_qwen_7b', 'statistical_only'")
    tokens_used = fields.Integer(help="if AI re-rank invoked")
    generation_ms = fields.Integer(help="latency for monitoring")

    # Lifecycle
    state = fields.Selection([
        ('pending', 'Pending review'),
        ('accepted', 'Accepted'),
        ('rejected', 'Rejected'),
        ('superseded', 'Superseded by newer suggestion'),
        ('stale', 'Stale (line changed since)'),
    ], default='pending', required=True, index=True)
    accepted_at = fields.Datetime()
    accepted_by = fields.Many2one('res.users')
    rejected_at = fields.Datetime()
    rejected_reason = fields.Selection([
        ('wrong_invoice', 'Wrong invoice'),
        ('wrong_partner', 'Wrong partner'),
        ('wrong_amount', 'Amount off'),
        ('not_a_match', 'No good match exists'),
        ('other', 'Other'),
    ])

    _confidence_in_range = models.Constraint(
        'CHECK (confidence >= 0.0 AND confidence <= 1.0)',
        'Confidence must be between 0.0 and 1.0',
    )

    @api.depends('confidence')
    def _compute_band(self):
        for sug in self:
            c = sug.confidence
            if c >= 0.95:
                sug.confidence_band = 'high'
            elif c >= 0.70:
                sug.confidence_band = 'medium'
            elif c >= 0.50:
                sug.confidence_band = 'low'
            else:
                sug.confidence_band = 'none'

    # ------------------------------------------------------------------
    # CRUD overrides — trigger MV refresh so the OWL widget sees fresh
    # confidence bands / top suggestion ids without waiting for cron.
    # ------------------------------------------------------------------

    @api.model_create_multi
    def create(self, vals_list):
        records = super().create(vals_list)
        self._trigger_mv_refresh()
        return records

    def write(self, vals):
        res = super().write(vals)
        # Only refresh on changes that affect the MV's projected columns.
        if 'state' in vals or 'confidence' in vals or 'rank' in vals:
            self._trigger_mv_refresh()
        return res

    def _trigger_mv_refresh(self):
        """Best-effort MV refresh; never poison the originating transaction.

        Uses concurrently=False because Postgres forbids
        REFRESH MATERIALIZED VIEW CONCURRENTLY inside a transaction block,
        and Odoo's per-request cursor is always in a transaction. The cron
        job (Task 25) opens a dedicated autocommit cursor for CONCURRENTLY
        refreshes when the MV grows large enough that a brief blocking
        refresh becomes objectionable.
        """
        try:
            self.env['fusion.unreconciled.bank.line.mv']._refresh(
                concurrently=False)
        except Exception as e:  # noqa: BLE001
            _logger.warning(
                "MV refresh after suggestion write failed: %s", e)