Add gtars genomicdist backend

bedboss/bedboss.py

@@ -434,9 +434,8 @@ def insert_pep(
     # Build the stats backend once for the whole batch. The backend holds
     # per-backend resources (persistent R service, gtars reference caches,
     # etc.) that should be reused across files.
-    stat_backend = (
-        build_backend(bbagent.config.config.analysis.backend) if not lite else None
-    )
+    backend_name = bbagent.config.config.analysis.backend
+    stat_backend = build_backend(backend_name) if not lite else None
 
     for i, pep_sample in enumerate(pep.samples):
         is_processed = skipper.is_processed(pep_sample.sample_name)
@@ -520,6 +519,7 @@ def insert_pep(
             force_overwrite=force_overwrite,
             annotation=bedset_annotation,
             lite=lite,
+            backend=backend_name,
         )
     else:
         _LOGGER.info(
@@ -786,4 +786,5 @@ def reprocess_bedset(
             )
         },
         lite=False,
+        backend=bbagent.config.config.analysis.backend,
     )

bedboss/bedbuncher/bedbuncher.py

@@ -102,6 +102,7 @@ def run_bedbuncher(
     no_fail: bool = False,
     force_overwrite: bool = False,
     lite: bool = False,
+    backend: str = "r",
 ) -> None:
     """
     Add bedset to the database
@@ -114,16 +115,17 @@ def run_bedbuncher(
     :param description: Bedset description
     :param annotation: Bedset annotation (author, source, summary, etc.)
     :param heavy: whether to use heavy processing (add all columns to the database).
-        if False -> R-script won't be executed, only basic statistics will be calculated
+        if False -> R-script won't be executed, only basic statistics will be calculated.
+        Ignored when backend is "gtars" (no R plots available).
     :param no_fail: whether to raise an error if bedset was not added to the database
     :param upload_pephub: whether to create a view in pephub
     :param upload_s3: whether to upload files to s3
     :param force_overwrite: whether to overwrite the record in the database
     :param lite: whether to run the pipeline in lite mode
-    # TODO: force_overwrite is not working!!! Fix it!
+    :param backend: analysis backend ("r" or "gtars"). For gtars, heavy is ignored.
     :return:
     """
-    _LOGGER.info(f"Adding bedset { record_id} to the database")
+    _LOGGER.info(f"Adding bedset {record_id} to the database")
 
     if isinstance(bedbase_config, str):
         bbagent = BedBaseAgent(bedbase_config)
@@ -140,15 +142,21 @@ def run_bedbuncher(
         "bedsets",
     )
 
-    if heavy:
+    gtars_like = backend == "gtars"
+    if heavy and not gtars_like:
         _LOGGER.info("Heavy processing is True. Calculating plots...")
         plot_value = create_plots(
             bedset=bed_set,
             output_folder=output_folder,
         )
         plots = BedSetPlots(region_commonality=FileModel(**plot_value))
     else:
-        _LOGGER.info("Heavy processing is False. Plots won't be calculated")
+        if gtars_like and heavy:
+            _LOGGER.info(
+                f"Heavy processing ignored for {backend} backend (no R plots available)"
+            )
+        else:
+            _LOGGER.info("Heavy processing is False. Plots won't be calculated")
         plots = None
 
     bbagent.bedset.create(
@@ -178,6 +186,7 @@ def run_bedbuncher_form_pep(
     upload_s3: bool = False,
     no_fail: bool = False,
     force_overwrite: bool = False,
+    backend: str = "r",
 ) -> str:
     """
     Create bedset from pep and add it to the database
@@ -224,6 +233,7 @@ def run_bedbuncher_form_pep(
         upload_s3=upload_s3,
         no_fail=no_fail,
         force_overwrite=force_overwrite,
+        backend=backend,
     )
     return bedset_name
 

bedboss/bedstat/backends/__init__.py

@@ -1,8 +1,15 @@
 from bedboss.bedstat.backends.base import StatBackend
+from bedboss.bedstat.backends.gtars_backend import GtarsStatBackend
 from bedboss.bedstat.backends.r_backend import RStatBackend
 from bedboss.const import BACKEND_GTARS, BACKEND_R
 
-__all__ = ["StatBackend", "RStatBackend", "create_backend", "build_backend"]
+__all__ = [
+    "StatBackend",
+    "RStatBackend",
+    "GtarsStatBackend",
+    "create_backend",
+    "build_backend",
+]
 
 
 def create_backend(name: str, **kwargs) -> StatBackend:
@@ -12,14 +19,14 @@ def create_backend(name: str, **kwargs) -> StatBackend:
     should use :func:`build_backend` instead, which handles backend-specific
     prerequisites (e.g. starting an RServiceManager for the R backend).
 
-    :param name: Backend name (BACKEND_R or BACKEND_GTARS)
+    :param name: Backend name ('r' or 'gtars')
     :param kwargs: Backend-specific keyword arguments
     :return: StatBackend instance
     """
     if name == BACKEND_R:
         return RStatBackend(**kwargs)
     elif name == BACKEND_GTARS:
-        raise NotImplementedError("gtars backend not yet available. Install via PR 2a.")
+        return GtarsStatBackend(**kwargs)
     else:
         raise ValueError(f"Unknown analysis backend: {name!r}. Use 'r' or 'gtars'.")
 
@@ -35,7 +42,7 @@ def build_backend(name: str) -> StatBackend:
     done to release backend-held resources. See `StatBackend` as a context
     manager for automatic cleanup.
 
-    :param name: Backend name (BACKEND_R or BACKEND_GTARS)
+    :param name: Backend name ('r' or 'gtars')
     :return: StatBackend instance ready for batch processing
     """
     if name == BACKEND_R: