Merge pull request #137 from databio/dev

Release 0.6.0

Dockerfile

@@ -27,7 +27,7 @@ RUN apt-get install -y python3-dev
 RUN apt-get install -y build-essential
 
 # Install CPU-only pytorch, eliminating huge nvidia dependencies
-RUN pip install torch==2.1.0+cpu -f https://download.pytorch.org/whl/torch_stable.html
+RUN pip install torch==2.3.1+cpu -f https://download.pytorch.org/whl/torch_stable.html
 RUN pip install https://github.com/pepkit/pipestat/archive/refs/heads/dev.zip
 
 RUN pip install -r requirements/requirements-all.txt --no-cache-dir

README.md

@@ -16,7 +16,7 @@ It needs a path to the *bedbase configuration file*, which can be provided eithe
 - **API**: <a href="https://api.bedbase.org/" target="_blank">https://api.bedbase.org/</a>
 - **DEV API**: <a href="https://dev.bedbase.org/" target="_blank">https://api-dev.bedbase.org/</a>
 - **UI**: <a href="https://bedbase.org/" target="_blank">https://bedbase.org/</a>
- -**DEV UI**: <a href="https://dev.bedhost.pages.dev/" target="_blank">https://dev.bedhost.pages.dev/</a>
+- **DEV UI**: <a href="https://dev.bedbase.org/" target="_blank">https://dev.bedbase.org/</a>
 - **Source Code**: <a href="https://github.com/databio/bedhost/" target="_blank">https://github.com/databio/bedhost/</a>
 
 ## Developer quick-start

bedhost/_version.py

@@ -1 +1 @@
-__version__ = "0.5.0"
+__version__ = "0.6.0"

bedhost/routers/base_api.py

@@ -13,12 +13,17 @@
 
 from ..main import bbagent, app
 from ..helpers import get_openapi_version
-from ..data_models import ServiceInfoResponse
+from ..data_models import (
+    ServiceInfoResponse,
+    Type,
+    Organization,
+    ComponentVersions,
+    EmbeddingModels,
+)
 from .._version import __version__ as bedhost_version
 
 router = APIRouter(prefix="/v1", tags=["base"])
 
-
 packages_versions = {}
 
 
@@ -31,7 +36,7 @@ async def get_bedbase_db_stats():
     """
     Returns statistics
     """
-    return bbagent.get_stats
+    return bbagent.get_stats()
 
 
 @router.get(
@@ -41,33 +46,33 @@ async def service_info():
     """
     Returns information about this service, such as versions, name, etc.
     """
-    all_versions = {
-        "bedhost_version": bedhost_version,
-        "bbconf_version": bbconf_version,
-        "geniml_version": geniml_version,
-        "python_version": python_version(),
-        "openapi_version": get_openapi_version(app),
-    }
+    all_versions = ComponentVersions(
+        bedhost_version=bedhost_version,
+        bbconf_version=bbconf_version,
+        geniml_version=geniml_version,
+        python_version=python_version(),
+        openapi_version=get_openapi_version(app),
+    )
 
     return ServiceInfoResponse(
         id="org.bedbase.api",
         name="BEDbase API",
-        type={
-            "group": "org.databio",
-            "artifact": "bedbase",
-            "version": bedhost_version,
-        },
+        type=Type(
+            group="org.databio",
+            artifact="bedbase",
+            version=bedhost_version,
+        ),
         description="An API providing genomic interval data and metadata",
-        organization={"name": "Databio Lab", "url": "https://databio.org"},
+        organization=Organization(name="Databio Lab", url="https://databio.org"),
         contactUrl="https://github.com/databio/bedbase/issues",
         documentationUrl="https://docs.bedbase.org",
-        updatedAt="2023-10-25T00:00:00Z",
+        updatedAt="2024-09-26T00:00:00Z",
         environment="main",
         version=bedhost_version,
         component_versions=all_versions,
-        embedding_models={
-            "vec2vec": bbagent.config.config.path.vec2vec,
-            "region2vec": bbagent.config.config.path.region2vec,
-            "text2vec": bbagent.config.config.path.text2vec,
-        },
+        embedding_models=EmbeddingModels(
+            vec2vec=bbagent.config.config.path.vec2vec,
+            region2vec=bbagent.config.config.path.region2vec,
+            text2vec=bbagent.config.config.path.text2vec,
+        ),
     )

bedhost/routers/bed_api.py

@@ -9,15 +9,15 @@
 from fastapi import APIRouter, HTTPException, Query, UploadFile, File
 from fastapi.responses import PlainTextResponse
 
-from genimtools.tokenizers import RegionSet
+from gtars.tokenizers import RegionSet
 
 import tempfile
 import os
 import shutil
 
 from bbconf.models.bed_models import (
     BedListResult,
-    BedMetadata,
+    BedMetadataAll,
     BedFiles,
     BedStatsModel,
     BedPlots,
@@ -45,7 +45,7 @@
 @router.get(
     "/example",
     summary="Get example BED record metadata",
-    response_model=BedMetadata,
+    response_model=BedMetadataAll,
 )
 async def get_example_bed_record():
     """
@@ -83,7 +83,8 @@ async def list_beds(
 @router.get(
     "/{bed_id}/metadata",
     summary="Get metadata for a single BED record",
-    response_model=BedMetadata,
+    response_model=BedMetadataAll,
+    response_model_by_alias=False,
     description=f"Example\n " f"bed_id: {EXAMPLE_BED}",
 )
 async def get_bed_metadata(
@@ -158,6 +159,7 @@ async def get_bed_stats(
     "/{bed_id}/metadata/classification",
     summary="Get classification of single BED file",
     response_model=BedClassification,
+    response_model_by_alias=False,
     description=f"Example\n bed_id: {EXAMPLE_BED}",
 )
 async def get_bed_classification(
@@ -176,6 +178,7 @@ async def get_bed_classification(
     summary="Get raw metadata for a single BED record",
     # response_model=BedPEPHub,
     response_model=BedPEPHubRestrict,
+    response_model_by_alias=False,
     description=f"Returns raw metadata for a single BED record. "
     f"This metadata is stored in PEPHub. And is not verified."
     f"Example\n bed_id: {EXAMPLE_BED}",
@@ -196,7 +199,7 @@ async def get_bed_pephub(
     summary="Get embeddings for a single BED record",
     response_model=BedEmbeddingResult,
 )
-def get_bed_embedding(bed_id: str = BedDigest):
+async def get_bed_embedding(bed_id: str = BedDigest):
     """
     Returns embeddings for a single BED record.
     """
@@ -208,6 +211,32 @@ def get_bed_embedding(bed_id: str = BedDigest):
         )
 
 
+@router.post(
+    "/embed",
+    summary="Get embeddings for a bed file.",
+    response_model=List[float],
+)
+async def embed_bed_file(
+    file: UploadFile = File(None),
+):
+    """
+    Create embedding for bed file
+    """
+    _LOGGER.info("Embedding file..")
+
+    if file is not None:
+        with tempfile.TemporaryDirectory() as dirpath:
+            file_path = os.path.join(dirpath, file.filename)
+
+            with open(file_path, "wb") as bed_file:
+                shutil.copyfileobj(file.file, bed_file)
+
+            region_set = RegionSet(file_path)
+
+            embedding = bbagent.bed._embed_file(region_set)
+    return embedding.tolist()[0]
+
+
 @router.get(
     "/{bed_id}/regions/{chr_num}",
     summary="Get regions from a BED file that overlap a query region.",
@@ -272,6 +301,7 @@ def get_regions_for_bedfile(
     summary="Search for a BedFile",
     tags=["search"],
     response_model=BedListSearchResult,
+    response_model_by_alias=False,
 )
 async def text_to_bed_search(query, limit: int = 10, offset: int = 0):
     """
@@ -291,6 +321,7 @@ async def text_to_bed_search(query, limit: int = 10, offset: int = 0):
     summary="Search for similar bed files",
     tags=["search"],
     response_model=BedListSearchResult,
+    response_model_by_alias=False,
 )
 async def bed_to_bed_search(
     file: UploadFile = File(None), limit: int = 10, offset: int = 0

bedhost/routers/bedset_api.py

@@ -23,6 +23,7 @@
     "/example",
     summary="Get metadata for an example BEDset record",
     response_model=BedSetMetadata,
+    response_model_by_alias=False,
 )
 async def get_example_bedset_record():
     result = bbagent.bedset.get_ids_list(limit=1).results
@@ -49,6 +50,7 @@ async def list_bedsets(query: str = None, limit: int = 1000, offset: int = 0):
     response_model=BedSetMetadata,
     summary="Get all metadata for a single BEDset record",
     description=f"Example\n bed_id: {EXAMPLE_BEDSET}",
+    response_model_by_alias=False,
 )
 async def get_bedset_metadata(
     bedset_id: str,
@@ -98,6 +100,7 @@ async def get_bedset_metadata(
     "/{bedset_id}/bedfiles",
     response_model=BedSetBedFiles,
     description=f"Example\n bed_id: {EXAMPLE_BEDSET}",
+    response_model_by_alias=False,
 )
 async def get_bedfiles_in_bedset(
     bedset_id: str,

dev.Dockerfile

@@ -27,7 +27,7 @@ RUN apt-get install -y python3-dev
 RUN apt-get install -y build-essential
 
 # Install CPU-only pytorch, eliminating huge nvidia dependencies
-RUN pip install torch==2.1.0+cpu -f https://download.pytorch.org/whl/torch_stable.html
+RUN pip install torch==2.3.1+cpu -f https://download.pytorch.org/whl/torch_stable.html
 RUN pip install https://github.com/pepkit/pipestat/archive/refs/heads/dev.zip
 
 RUN pip install -r requirements/requirements-all.txt --no-cache-dir

requirements/requirements-all.txt

@@ -1,5 +1,5 @@
-bbconf @ git+https://github.com/databio/bbconf.git@dev#egg=bbconf
-# bbconf>=0.6.0
+#bbconf @ git+https://github.com/databio/bbconf.git@dev#egg=bbconf
+bbconf>=0.7.1
 fastapi>=0.103.0
 logmuse>=0.2.7
 markdown