Bacpop-207 Add info_csv file for metadata

beebop/app.py

@@ -11,12 +11,15 @@
 import json
 import requests
 import pickle
+from datetime import datetime
+from typing import Any
 
 from beebop import versions, assignClusters, visualise
 from beebop.filestore import PoppunkFileStore, DatabaseFileStore
 from beebop.utils import get_args, get_cluster_num
 from PopPUNK.sketchlib import getKmersFromReferenceDatabase
 import beebop.schemas
+from beebop.dataClasses import SpeciesConfig
 schemas = beebop.schemas.Schema()
 
 redis_host = os.environ.get("REDIS_HOST")
@@ -326,7 +329,7 @@ def run_poppunk_internal(sketches: dict,
 
 
 def setup_db_file_stores(
-    species_args: dict,
+    species_args: SpeciesConfig,
 ) -> tuple[DatabaseFileStore, DatabaseFileStore]:
     """
     [Initializes the reference and full database file stores
@@ -340,12 +343,14 @@ def setup_db_file_stores(
     ref_db_fs = DatabaseFileStore(
         f"{dbs_location}/{species_args.refdb}",
         species_args.external_clusters_file,
+        species_args.db_metadata_file,
     )
 
     if os.path.exists(f"{dbs_location}/{species_args.fulldb}"):
         full_db_fs = DatabaseFileStore(
             f"{dbs_location}/{species_args.fulldb}",
             species_args.external_clusters_file,
+            species_args.db_metadata_file,
         )
     else:
         full_db_fs = ref_db_fs
@@ -595,6 +600,9 @@ def generate_microreact_url_internal(microreact_api_new_url: str,
     with open(path_json, 'rb') as microreact_file:
         json_microreact = json.load(microreact_file)
 
+    json_microreact["meta"][
+        "name"
+    ] = f"Cluster {cluster_num} - {datetime.now().strftime('%Y-%m-%d %H:%M')}"
     # generate URL from microreact API
     headers = {"Content-type": "application/json; charset=UTF-8",
                "Access-Token": api_token}

beebop/assignClusters.py

@@ -11,8 +11,7 @@
 from beebop.poppunkWrapper import PoppunkWrapper
 from beebop.filestore import PoppunkFileStore, DatabaseFileStore
 import shutil
-from typing import Optional, Any
-from dataclasses import dataclass
+from beebop.dataClasses import ClusteringConfig
 
 
 def hex_to_decimal(sketches_dict) -> None:
@@ -33,19 +32,6 @@ def hex_to_decimal(sketches_dict) -> None:
                 sample[key] = list(map(lambda x: int(x, 16), value))
 
 
-@dataclass
-class ClusteringConfig:
-    species: str
-    p_hash: str
-    args: dict
-    external_clusters_prefix: Optional[str]
-    fs: PoppunkFileStore
-    full_db_fs: DatabaseFileStore
-    ref_db_fs: DatabaseFileStore
-    db_funcs: dict[str, Any]
-    out_dir: str
-
-
 def get_clusters(
     hashes_list: list,
     p_hash: str,

beebop/dataClasses.py

@@ -0,0 +1,42 @@
+from dataclasses import dataclass
+from typing import Optional, Any
+from beebop.filestore import PoppunkFileStore, DatabaseFileStore
+
+
+@dataclass
+class ClusteringConfig:
+    species: str
+    p_hash: str
+    args: dict
+    external_clusters_prefix: Optional[str]
+    fs: PoppunkFileStore
+    full_db_fs: DatabaseFileStore
+    ref_db_fs: DatabaseFileStore
+    db_funcs: dict[str, Any]
+    out_dir: str
+
+
+@dataclass
+class Qc:
+    run_qc: bool
+    type_isolate: Optional[str]
+    max_a_dist: float
+    max_pi_dist: float
+    prop_zero: float
+    prop_n: float
+    max_merge: int
+    betweenness: bool
+    retain_failures: bool
+    no_remove: bool
+    length_range: list[int]
+    upper_n: Optional[int]
+
+
+@dataclass
+class SpeciesConfig:
+    refdb: str
+    fulldb: str
+    external_cluster_prefix: str
+    external_clusters_file: str
+    db_metadata_file: str
+    qc_dict: Qc

beebop/filestore.py

@@ -292,7 +292,10 @@ class DatabaseFileStore:
     """
 
     def __init__(
-        self, full_path: str, external_clusters_file: Optional[str] = None
+        self,
+        full_path: str,
+        external_clusters_file: Optional[str] = None,
+        db_metadata_file: Optional[str] = None,
     ):
         """
         :param full_path: [path to database]
@@ -311,3 +314,8 @@ def __init__(
             if external_clusters_file
             else None
         )
+        self.metadata = (
+            str(PurePath("beebop", "resources", db_metadata_file))
+            if db_metadata_file
+            else None
+        )

beebop/poppunkWrapper.py

@@ -110,8 +110,8 @@ def create_microreact(self, cluster: str, internal_cluster: str) -> None:
             previous_distances=None,
             network_file=None,
             gpu_graph=self.args.visualise.gpu_graph,
-            info_csv=self.args.visualise.info_csv,
-            rapidnj=shutil.which("rapidnj"),
+            info_csv=self.db_fs.metadata,
+            rapidnj=shutil.which('rapidnj'),
             api_key=None,
             tree=self.args.visualise.tree,
             mst_distances=self.args.visualise.mst_distances,