feat: implement LLM table translation service with searchable datasource dropdown, i18n, sidebar, and RBAC

- Add backend plugin modules: preview, executor, orchestrator, events, sql_generator, superset_executor, dictionary, scheduler, metrics
- Add API routes for translate jobs, runs, dictionaries, preview, schedule, metrics, corrections
- Add ORM models (12 tables) and Pydantic schemas (15 DTOs)
- Register translate router in app.py with RBAC permission guards
- Add frontend pages: job list, job config, dictionary list/editor, history
- Add 7 reusable Svelte 5 components: Preview, RunProgress, RunResult, TermCorrection, BulkCorrection, ScheduleConfig, MetricsDashboard
- Add searchable datasource dropdown with Superset API integration
- Add i18n (en/ru) with ~300 keys across common, config, jobs, dictionaries, preview namespaces
- Add Translation sidebar category with Jobs/Dictionaries/History sub-items
- Hide health monitor error toast via suppressToast API option
- Add 69 backend tests and 44 frontend test files
- Fix: SupersetClient env resolution (string -> Environment object)
- Fix: Dataset detail API returns proper database dict
- Fix: Database dialect extraction fallback when metadata incomplete
This commit is contained in:
2026-05-09 19:34:25 +03:00
parent bf82e17418
commit 67ba04d4ff
44 changed files with 14744 additions and 5 deletions

View File

@@ -0,0 +1,292 @@
# [DEF:TranslateModels:Module]
# @COMPLEXITY: 3
# @SEMANTICS: translate, models, sqlalchemy, persistence
# @PURPOSE: SQLAlchemy ORM models for LLM-based SQL/dashboard translation across dialects.
# @LAYER: Domain
# @RELATION: INHERITS_FROM -> MappingModels:Base
from sqlalchemy import (
Column, String, Boolean, DateTime, JSON, Text,
ForeignKey, Integer, UniqueConstraint, Index
)
from sqlalchemy.dialects.postgresql import JSON as PG_JSON
from datetime import datetime, timezone
import uuid
import hashlib
from .mapping import Base
def generate_uuid():
return str(uuid.uuid4())
# [DEF:TranslationJob:Class]
# @PURPOSE: A translation job representing a multi-dialect conversion task with column mappings, LLM config, and dictionary attachments.
# @RATIONALE: Snapshot isolation — in-progress runs use config snapshot; config edits affect future runs only.
# @REJECTED: Invalidating in-progress runs on config edit would break scheduled run continuity.
class TranslationJob(Base):
__tablename__ = "translation_jobs"
id = Column(String, primary_key=True, default=generate_uuid)
name = Column(String, nullable=False)
description = Column(Text, nullable=True)
source_dialect = Column(String, nullable=False)
target_dialect = Column(String, nullable=False)
database_dialect = Column(String, nullable=True, comment="Detected dialect from Superset connection at save time")
status = Column(String, nullable=False, default="DRAFT") # DRAFT, READY, RUNNING, COMPLETED, FAILED, CANCELLED
# Datasource & target table configuration
source_datasource_id = Column(String, nullable=True, comment="Superset datasource ID")
source_table = Column(String, nullable=True, comment="Source table name resolved from datasource")
target_schema = Column(String, nullable=True, comment="Target table schema")
target_table = Column(String, nullable=True, comment="Target table name")
# Column mapping
source_key_cols = Column(JSON, nullable=True, comment="Source key column names for composite key")
target_key_cols = Column(JSON, nullable=True, comment="Target key column names for composite key")
translation_column = Column(String, nullable=True, comment="The column whose values will be translated")
context_columns = Column(JSON, nullable=True, comment="Context column names included in LLM prompt")
# LLM & processing settings
target_language = Column(String, nullable=True, comment="Target language code (e.g. en, ru)")
provider_id = Column(String, nullable=True, comment="LLM provider ID")
batch_size = Column(Integer, nullable=False, default=50, comment="Records per batch")
upsert_strategy = Column(String, nullable=False, default="MERGE", comment="MERGE, INSERT, UPDATE")
created_by = Column(String, nullable=True)
created_at = Column(DateTime, default=lambda: datetime.now(timezone.utc))
updated_at = Column(DateTime, default=lambda: datetime.now(timezone.utc), onupdate=lambda: datetime.now(timezone.utc))
# [/DEF:TranslationJob:Class]
# [DEF:TranslationRun:Class]
# @PURPOSE: Represents a single execution of a translation job, capturing status, timing, and Superset execution metadata.
class TranslationRun(Base):
__tablename__ = "translation_runs"
id = Column(String, primary_key=True, default=generate_uuid)
job_id = Column(String, ForeignKey("translation_jobs.id"), nullable=False, index=True)
status = Column(String, nullable=False, default="PENDING") # PENDING, RUNNING, COMPLETED, FAILED, CANCELLED
trigger_type = Column(String, nullable=True, comment="manual, scheduled, retry, baseline_expired")
started_at = Column(DateTime, nullable=True)
completed_at = Column(DateTime, nullable=True)
error_message = Column(Text, nullable=True)
total_records = Column(Integer, default=0)
successful_records = Column(Integer, default=0)
failed_records = Column(Integer, default=0)
skipped_records = Column(Integer, default=0)
insert_status = Column(String, nullable=True, comment="Status of the Superset insert/update operation")
superset_execution_id = Column(String, nullable=True, comment="Superset execution/task ID")
superset_execution_log = Column(JSON, nullable=True, comment="Superset execution log output")
config_snapshot = Column(JSON, nullable=True, comment="Snapshot of job config at run creation time")
key_hash = Column(String, nullable=True, comment="Hash of source key fields for dedup")
config_hash = Column(String, nullable=True, comment="Hash of translation configuration state")
dict_snapshot_hash = Column(String, nullable=True, comment="Hash of dictionary state at run time")
created_by = Column(String, nullable=True)
created_at = Column(DateTime, default=lambda: datetime.now(timezone.utc))
# [/DEF:TranslationRun:Class]
# [DEF:TranslationBatch:Class]
# @PURPOSE: Groups translation records within a run into manageable batches with timing and record counts.
class TranslationBatch(Base):
__tablename__ = "translation_batches"
id = Column(String, primary_key=True, default=generate_uuid)
run_id = Column(String, ForeignKey("translation_runs.id"), nullable=False, index=True)
batch_index = Column(Integer, nullable=False)
status = Column(String, nullable=False, default="PENDING")
total_records = Column(Integer, default=0)
successful_records = Column(Integer, default=0)
failed_records = Column(Integer, default=0)
started_at = Column(DateTime, nullable=True)
completed_at = Column(DateTime, nullable=True)
created_at = Column(DateTime, default=lambda: datetime.now(timezone.utc))
# [/DEF:TranslationBatch:Class]
# [DEF:TranslationRecord:Class]
# @PURPOSE: Individual translation result for a single SQL statement or dashboard element, tracking source, target, and error state.
class TranslationRecord(Base):
__tablename__ = "translation_records"
id = Column(String, primary_key=True, default=generate_uuid)
batch_id = Column(String, ForeignKey("translation_batches.id"), nullable=False, index=True)
run_id = Column(String, ForeignKey("translation_runs.id"), nullable=False, index=True)
source_sql = Column(Text, nullable=True)
target_sql = Column(Text, nullable=True)
source_object_type = Column(String, nullable=True) # query, dashboard, chart, dataset
source_object_id = Column(String, nullable=True)
source_object_name = Column(String, nullable=True)
status = Column(String, nullable=False, default="PENDING") # PENDING, SUCCESS, FAILED, SKIPPED
error_message = Column(Text, nullable=True)
token_count_input = Column(Integer, nullable=True)
token_count_output = Column(Integer, nullable=True)
translation_duration_ms = Column(Integer, nullable=True)
created_at = Column(DateTime, default=lambda: datetime.now(timezone.utc))
__table_args__ = (
Index("ix_translation_records_run_status", "run_id", "status"),
)
# [/DEF:TranslationRecord:Class]
# [DEF:TranslationEvent:Class]
# @PURPOSE: Audit/event log for translation operations, with optional run_id for context.
class TranslationEvent(Base):
__tablename__ = "translation_events"
id = Column(String, primary_key=True, default=generate_uuid)
job_id = Column(String, ForeignKey("translation_jobs.id"), nullable=False, index=True)
run_id = Column(String, ForeignKey("translation_runs.id"), nullable=True, index=True)
event_type = Column(String, nullable=False) # JOB_CREATED, JOB_STARTED, JOB_COMPLETED, JOB_FAILED, etc.
event_data = Column(JSON, nullable=True)
created_by = Column(String, nullable=True)
created_at = Column(DateTime, default=lambda: datetime.now(timezone.utc))
# [/DEF:TranslationEvent:Class]
# [DEF:TranslationPreviewSession:Class]
# @PURPOSE: A preview session allowing users to review proposed translations before applying them.
class TranslationPreviewSession(Base):
__tablename__ = "translation_preview_sessions"
id = Column(String, primary_key=True, default=generate_uuid)
job_id = Column(String, ForeignKey("translation_jobs.id"), nullable=False, index=True)
run_id = Column(String, ForeignKey("translation_runs.id"), nullable=True)
status = Column(String, nullable=False, default="ACTIVE") # ACTIVE, APPLIED, DISCARDED, EXPIRED
created_by = Column(String, nullable=True)
created_at = Column(DateTime, default=lambda: datetime.now(timezone.utc))
expires_at = Column(DateTime, nullable=True)
# [/DEF:TranslationPreviewSession:Class]
# [DEF:TranslationPreviewRecord:Class]
# @PURPOSE: Individual preview entry within a preview session, showing original and translated content side by side.
class TranslationPreviewRecord(Base):
__tablename__ = "translation_preview_records"
id = Column(String, primary_key=True, default=generate_uuid)
session_id = Column(String, ForeignKey("translation_preview_sessions.id"), nullable=False, index=True)
source_sql = Column(Text, nullable=True)
target_sql = Column(Text, nullable=True)
source_object_type = Column(String, nullable=True)
source_object_id = Column(String, nullable=True)
source_object_name = Column(String, nullable=True)
status = Column(String, nullable=False, default="PENDING") # PENDING, APPROVED, REJECTED
feedback = Column(Text, nullable=True)
created_at = Column(DateTime, default=lambda: datetime.now(timezone.utc))
# [/DEF:TranslationPreviewRecord:Class]
# [DEF:TerminologyDictionary:Class]
# @PURPOSE: A named collection of terminology mappings used during translation to ensure consistent term translation.
class TerminologyDictionary(Base):
__tablename__ = "terminology_dictionaries"
id = Column(String, primary_key=True, default=generate_uuid)
name = Column(String, nullable=False)
description = Column(Text, nullable=True)
source_dialect = Column(String, nullable=False)
target_dialect = Column(String, nullable=False)
is_active = Column(Boolean, default=True)
created_by = Column(String, nullable=True)
created_at = Column(DateTime, default=lambda: datetime.now(timezone.utc))
updated_at = Column(DateTime, default=lambda: datetime.now(timezone.utc), onupdate=lambda: datetime.now(timezone.utc))
# [/DEF:TerminologyDictionary:Class]
# [DEF:DictionaryEntry:Class]
# @PURPOSE: A single term mapping entry with case-insensitive unique constraint on source_term_normalized and origin tracking.
class DictionaryEntry(Base):
__tablename__ = "dictionary_entries"
id = Column(String, primary_key=True, default=generate_uuid)
dictionary_id = Column(String, ForeignKey("terminology_dictionaries.id"), nullable=False, index=True)
source_term = Column(String, nullable=False)
source_term_normalized = Column(String, nullable=False)
target_term = Column(String, nullable=False)
context_notes = Column(Text, nullable=True)
origin_run_id = Column(String, nullable=True, comment="Run ID from which this correction originated")
origin_row_key = Column(String, nullable=True, comment="Row key within the run that triggered this correction")
origin_user_id = Column(String, nullable=True, comment="User who submitted the correction")
created_at = Column(DateTime, default=lambda: datetime.now(timezone.utc))
updated_at = Column(DateTime, default=lambda: datetime.now(timezone.utc), onupdate=lambda: datetime.now(timezone.utc))
__table_args__ = (
UniqueConstraint(
"dictionary_id", "source_term_normalized",
name="uq_dictionary_entry_term"
),
)
# [/DEF:DictionaryEntry:Class]
# [DEF:TranslationSchedule:Class]
# @PURPOSE: Defines a cron-based schedule for recurring translation jobs.
class TranslationSchedule(Base):
__tablename__ = "translation_schedules"
id = Column(String, primary_key=True, default=generate_uuid)
job_id = Column(String, ForeignKey("translation_jobs.id"), nullable=False, index=True)
cron_expression = Column(String, nullable=False)
timezone = Column(String, nullable=False, default="UTC")
is_active = Column(Boolean, default=True)
last_run_at = Column(DateTime, nullable=True)
next_run_at = Column(DateTime, nullable=True)
created_by = Column(String, nullable=True)
created_at = Column(DateTime, default=lambda: datetime.now(timezone.utc))
updated_at = Column(DateTime, default=lambda: datetime.now(timezone.utc), onupdate=lambda: datetime.now(timezone.utc))
# [/DEF:TranslationSchedule:Class]
# [DEF:TranslationJobDictionary:Class]
# @PURPOSE: Many-to-many association between translation jobs and terminology dictionaries.
class TranslationJobDictionary(Base):
__tablename__ = "translation_job_dictionaries"
id = Column(String, primary_key=True, default=generate_uuid)
job_id = Column(String, ForeignKey("translation_jobs.id"), nullable=False, index=True)
dictionary_id = Column(String, ForeignKey("terminology_dictionaries.id"), nullable=False, index=True)
created_at = Column(DateTime, default=lambda: datetime.now(timezone.utc))
__table_args__ = (
UniqueConstraint(
"job_id", "dictionary_id",
name="uq_job_dictionary"
),
)
# [/DEF:TranslationJobDictionary:Class]
# [DEF:MetricSnapshot:Class]
# @PURPOSE: Captures translation performance metrics at a point in time with config/content/dictionary hash keys for aggregation.
class MetricSnapshot(Base):
__tablename__ = "translation_metric_snapshots"
id = Column(String, primary_key=True, default=generate_uuid)
job_id = Column(String, ForeignKey("translation_jobs.id"), nullable=False, index=True)
run_id = Column(String, ForeignKey("translation_runs.id"), nullable=True, index=True)
key_hash = Column(String, nullable=False, comment="Hash of dimension key fields for aggregation")
config_hash = Column(String, nullable=True, comment="Hash of translation configuration state")
dict_snapshot_hash = Column(String, nullable=True, comment="Hash of dictionary state at capture time")
covers_events_before = Column(DateTime, nullable=True, comment="Indicates snapshot covers events before this timestamp")
total_jobs = Column(Integer, default=0)
total_runs = Column(Integer, default=0)
total_records = Column(Integer, default=0)
successful_records = Column(Integer, default=0)
failed_records = Column(Integer, default=0)
skipped_records = Column(Integer, default=0)
avg_duration_ms = Column(Integer, nullable=True)
p50_duration_ms = Column(Integer, nullable=True)
p95_duration_ms = Column(Integer, nullable=True)
p99_duration_ms = Column(Integer, nullable=True)
snapshot_date = Column(DateTime, nullable=False, default=lambda: datetime.now(timezone.utc))
created_at = Column(DateTime, default=lambda: datetime.now(timezone.utc))
__table_args__ = (
Index("ix_metric_snapshots_job_date", "job_id", "snapshot_date"),
)
# [/DEF:MetricSnapshot:Class]
# [/DEF:TranslateModels:Module]