Make db optional (#77)

### Added

Make databases optional with --no-cosmic --no-fusiongdb2 --no-mitelman options

CHANGELOG.md

@@ -5,6 +5,12 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [3.0.0]
+
+### Added
+
+- Options --no-cosmic/--no-fusiongdb2/--no-mitelman to download and run without those specified databases
+
 ## [2.1.8]
 
 ### Removed

docs/download.md

@@ -25,6 +25,14 @@ fusion_report download
     /path/to/db
 ```
 
+You can exclude a specific database with --no-cosmic/--no-mitelman/--no-fusiongdb2. Exemple for no cosmic:
+
+```bash
+fusion_report download
+    --no-cosmic
+    /path/to/db
+```
+
 
 ## Manual download
 

fusion_report/app.py

@@ -8,7 +8,7 @@
 from collections import defaultdict
 from typing import Any, Dict, List
 
-import rapidjson
+import json
 
 from tqdm import tqdm
 
@@ -81,7 +81,7 @@ def run(self):
     def preprocess(self, params: Namespace) -> None:
         """Parse, enrich and score fusion."""
         self.parse_fusion_outputs(vars(params))
-        self.enrich(params.db_path)
+        self.enrich(params)
         self.score(vars(params))
 
     def generate_report(self, params: Namespace) -> None:
@@ -118,13 +118,25 @@ def parse_fusion_outputs(self, params: Dict[str, Any]) -> None:
                 # value: fusion tool output
                 self.manager.parse(param, value, params["allow_multiple_gene_symbols"])
 
-    def enrich(self, path: str) -> None:
+    def enrich(self, params) -> None:
         """Enrich fusion with all relevant information from local databases."""
-        local_fusions: Dict[str, List[str]] = {
-            CosmicDB(path).name: CosmicDB(path).get_all_fusions(),
-            MitelmanDB(path).name: MitelmanDB(path).get_all_fusions(),
-            FusionGDB2(path).name: FusionGDB2(path).get_all_fusions(),
-        }
+        local_fusions: Dict[str, List[str]] = {}
+
+        if not params.no_cosmic:
+            local_fusions.update(
+                {CosmicDB(params.db_path).name: CosmicDB(params.db_path).get_all_fusions()}
+            )
+
+        if not params.no_fusiongdb2:
+            local_fusions.update(
+                {MitelmanDB(params.db_path).name: MitelmanDB(params.db_path).get_all_fusions()}
+            )
+
+        if not params.no_mitelman:
+            local_fusions.update(
+                {FusionGDB2(params.db_path).name: FusionGDB2(params.db_path).get_all_fusions()}
+            )
+
         for fusion in self.manager.fusions:
             for db_name, db_list in local_fusions.items():
                 if fusion.name in db_list:
@@ -138,7 +150,7 @@ def export_results(self, path: str, extension: str) -> None:
         if extension == "json":
             with open(dest, "w", encoding="utf-8") as output:
                 results = [fusion.json_serialize() for fusion in self.manager.fusions]
-                output.write(rapidjson.dumps(results))
+                output.write(json.dumps(results))
         elif extension == "csv":
             with open(dest, "w", encoding="utf-8") as output:
                 csv_writer = csv.writer(
@@ -248,4 +260,4 @@ def generate_multiqc(
 
         dest = f"{os.path.join(path, 'fusion_genes_mqc.json')}"
         with open(dest, "w", encoding="utf-8") as output:
-            output.write(rapidjson.dumps(configuration))
+            output.write(json.dumps(configuration))

fusion_report/args_builder.py

@@ -4,7 +4,7 @@
 from argparse import ArgumentParser, Namespace, _SubParsersAction
 from typing import Any, Dict
 
-import rapidjson
+import json
 
 from fusion_report.settings import Settings
 
@@ -21,7 +21,7 @@ class ArgsBuilder:
     def __init__(self):
         configuration = os.path.join(Settings.ROOT_DIR, "arguments.json")
         with open(configuration, "r") as config_file:
-            self.arguments: Dict[str, Any] = rapidjson.loads(config_file.read())
+            self.arguments: Dict[str, Any] = json.load(config_file)
         self.arguments["weight"] = float(100 / len(self.supported_tools))
         self.parser = ArgumentParser(
             description="""Tool for generating friendly UI custom report."""
@@ -100,6 +100,13 @@ def run_args(self, args, weight) -> None:
                         type=type(optional.get("default")),
                     )
 
+        for database in args["databases"]:
+            run_parser.add_argument(
+                database["key"],
+                help=database["help"],
+                action=database.get("action", "store"),
+            )
+
     def download_args(self, args: Dict[str, Any]) -> None:
         """Build download command-line arguments."""
         download_parser = self.command_parser.add_parser(
@@ -108,6 +115,13 @@ def download_args(self, args: Dict[str, Any]) -> None:
         for mandatory in args["mandatory"]:
             download_parser.add_argument(mandatory["key"], help=mandatory["help"], type=str)
 
+        for optional in args["optionals"]:
+            download_parser.add_argument(
+                optional["key"],
+                help=optional["help"],
+                action=optional.get("action", "store"),
+            )
+
         self._cosmic(args, download_parser)
 
     def sync_args(self, args: Dict[str, Any]) -> None:

fusion_report/arguments.json

@@ -49,6 +49,23 @@
                     "help": "Jaffa output file"
                 }
             ],
+            "databases": [
+                {
+                    "key": "--no-cosmic",
+                    "help": "Do not download cosmic fusion database",
+                    "action": "store_true"
+                },
+                {
+                    "key": "--no-fusiongdb2",
+                    "help": "Do not download fusiongdb2 fusion database",
+                    "action": "store_true"
+                },
+                {
+                    "key": "--no-mitelman",
+                    "help": "Do not download mitelman fusion database",
+                    "action": "store_true"
+                }
+            ],
             "optionals": [
                 {
                     "key": ["--allow-multiple-gene-symbols"],
@@ -79,6 +96,23 @@
                     "help": "Output directory"
                 }
             ],
+            "optionals": [
+                {
+                    "key": "--no-cosmic",
+                    "help": "Do not download cosmic fusion database",
+                    "action": "store_true"
+                },
+                {
+                    "key": "--no-fusiongdb2",
+                    "help": "Do not download fusiongdb2 fusion database",
+                    "action": "store_true"
+                },
+                {
+                    "key": "--no-mitelman",
+                    "help": "Do not download mitelman fusion database",
+                    "action": "store_true"
+                }
+            ],
             "cosmic": [
                 {
                     "key": "--cosmic_usr",

fusion_report/download.py

@@ -19,7 +19,6 @@ class Download:
     """
 
     def __init__(self, params: Namespace):
-        self.validate(params)
         self.download_all(params)
 
     def validate(self, params: Namespace) -> None:
@@ -29,28 +28,32 @@ def validate(self, params: Namespace) -> None:
         else:
             self.cosmic_token = Net.get_cosmic_token(params)
 
+    def download_all(self, params: Namespace) -> None:
         # making sure output directory exists
         if not os.path.exists(params.output):
             os.makedirs(params.output, 0o755)
 
-    def download_all(self, params: Namespace) -> None:
         """Download all databases."""
         return_err: List[str] = []
         os.chdir(params.output)
 
-        # MITELMAN
-        Net.get_mitelman(self, return_err)
+        if not params.no_mitelman:
+            # MITELMAN
+            Net.get_mitelman(self, return_err)
 
-        # FusionGDB2
-        Net.get_fusiongdb2(self, return_err)
+        if not params.no_fusiongdb2:
+            # FusionGDB2
+            Net.get_fusiongdb2(self, return_err)
 
-        # COSMIC
-        if params.qiagen:
-            Logger(__name__).info("Downloading resources from QIAGEN...")
-            Net.get_cosmic_from_qiagen(self.cosmic_token, return_err, params.output)
-        else:
-            Logger(__name__).info("Downloading resources from SANGER...")
-            Net.get_cosmic_from_sanger(self.cosmic_token, return_err)
+        if not params.no_cosmic:
+            # COSMIC
+            self.validate(params)
+            if params.qiagen:
+                Logger(__name__).info("Downloading resources from QIAGEN...")
+                Net.get_cosmic_from_qiagen(self.cosmic_token, return_err, params.output)
+            else:
+                Logger(__name__).info("Downloading resources from SANGER...")
+                Net.get_cosmic_from_sanger(self.cosmic_token, return_err)
 
         if len(return_err) > 0:
             raise DownloadException(return_err)

requirements.txt

@@ -4,7 +4,6 @@ pyyaml>=4.2b1
 Jinja2>=2.10
 MarkupSafe>=2.1.1
 pandas
-python-rapidjson
 requests
 openpyxl
 xlrd >= 2.0.0

setup.cfg

@@ -6,4 +6,4 @@ lines_after_imports=2
 lines_between_types=1
 multi_line_output=7
 use_parentheses=true
-known_third_party=colorlog,jinja2,rapidjson,setuptools,tqdm,yaml
+known_third_party=colorlog,jinja2,setuptools,tqdm,yaml