update cli for serve/process, add test

Author: Jhsmit

pyhdx/batch_processing.py

@@ -237,7 +237,7 @@ class JobParser(object):
 
     cwd = param.ClassSelector(Path, doc='Path of the current working directory')
 
-    def __init__(self, job_spec, cwd=None, ):
+    def __init__(self, job_spec, cwd=None):
         self.job_spec = job_spec
         self.cwd = cwd or Path().cwd()
 

pyhdx/cli.py

@@ -1,32 +1,30 @@
-import argparse
 import time
-from ipaddress import ip_address
-from pyhdx.web import serve
-from pyhdx.config import cfg
-from pyhdx.local_cluster import verify_cluster, default_cluster
+from typing import Union, Optional
+from pathlib import Path
 
+import typer
+from ipaddress import ip_address
+import yaml
 
-# todo add check to see if the web module requirements are installed
 
+app = typer.Typer()
 
-def main():
-    parser = argparse.ArgumentParser(prog="pyhdx", description="PyHDX Launcher")
+@app.command()
+def serve(scheduler_address: Optional[str] = typer.Option(None, help="Address for dask scheduler to use")):
+    """Launch the PyHDX web application"""
 
-    parser.add_argument("serve", help="Runs PyHDX Dashboard")
-    parser.add_argument(
-        "--scheduler_address", help="Run with local cluster <ip>:<port>"
-    )
-    args = parser.parse_args()
+    from pyhdx.config import cfg
+    from pyhdx.local_cluster import verify_cluster, default_cluster
 
-    if args.scheduler_address:
-        ip, port = args.scheduler_address.split(":")
+    if scheduler_address is not None:
+        ip, port = scheduler_address.split(":")
         if not ip_address(ip):
             print("Invalid IP Address")
             return
         elif not 0 <= int(port) < 2 ** 16:
             print("Invalid port, must be 0-65535")
             return
-        cfg.set("cluster", "scheduler_address", args.scheduler_address)
+        cfg.set("cluster", "scheduler_address", scheduler_address)
 
     scheduler_address = cfg.get("cluster", "scheduler_address")
     if not verify_cluster(scheduler_address):
@@ -37,8 +35,9 @@ def main():
         scheduler_address = f"{ip}:{port}"
         print(f"Started new Dask LocalCluster at {scheduler_address}")
 
-    if args.serve:
-        serve.run_apps()
+    # Start the PyHDX web application
+    from pyhdx.web import serve as serve_pyhdx
+    serve_pyhdx.run_apps()
 
     loop = True
     while loop:
@@ -49,11 +48,22 @@ def main():
             loop = False
 
 
-if __name__ == "__main__":
-    import sys
+@app.command()
+def process(
+        jobfile: Path = typer.Argument(..., help="Path to .yaml jobfile"),
+        cwd: Optional[Path] = typer.Option(None, help="Optional path to working directory")
+):
+    """
+    Process a HDX dataset according to a jobfile
+    """
+
+    from pyhdx.batch_processing import JobParser
 
-    sys.argv.append("serve")
-    sys.argv.append("--scheduler_address")
-    sys.argv.append("127.0.0.1:53270")
+    job_spec = yaml.safe_load(jobfile.read_text())
+    parser = JobParser(job_spec, cwd=cwd)
 
-    main()
+    parser.execute()
+
+
+if __name__ == "__main__":
+    app()

setup.cfg

@@ -1,9 +1,9 @@
 [metadata]
 name = PyHDX
 author = Jochem H. Smit
-author-email = [email protected]
+author_email = [email protected]
 maintainer = Jochem H. Smit
-maintainer-email = [email protected]
+maintainer_email = [email protected]
 url = https://github.com/Jhsmit/PyHDX
 license = MIT
 license_files = LICENSE
@@ -31,6 +31,7 @@ install_requires =
     sympy==1.5.1
     torch
     tqdm
+    typer
     dask
     distributed
     packaging
@@ -41,7 +42,7 @@ python_requires =
 
 [options.entry_points]
 console_scripts =
-    pyhdx = pyhdx.cli:main
+    pyhdx = pyhdx.cli:app
 
 
 [options.extras_require]

tests/test_batchprocessing.py

@@ -1,8 +1,9 @@
-from pyhdx.batch_processing import yaml_to_hdxm, yaml_to_hdxmset, StateParser
+from pyhdx.batch_processing import StateParser, JobParser
 from pyhdx.models import HDXMeasurement, HDXMeasurementSet
 import numpy as np
 from pathlib import Path
 import yaml
+import shutil
 
 cwd = Path(__file__).parent
 input_dir = cwd / 'test_data' / 'input'
@@ -17,16 +18,6 @@ def test_load_from_yaml(self):
         yaml_pth = Path(input_dir / 'data_states.yaml')
         data_dict = yaml.safe_load(yaml_pth.read_text())
 
-        hdxm = yaml_to_hdxm(data_dict['SecB_tetramer'], data_dir=input_dir)
-        assert isinstance(hdxm, HDXMeasurement)
-
-        assert hdxm.metadata['temperature'] == data_dict['SecB_tetramer']['temperature']['value'] + 273.15
-        assert hdxm.name == 'SecB WT apo'
-
-        hdxm_set = yaml_to_hdxmset(data_dict, data_dir=input_dir)
-        assert isinstance(hdxm_set, HDXMeasurementSet)
-        assert hdxm_set.names == list(data_dict.keys())
-
         parser = StateParser(data_dict, data_src=input_dir)
 
         hdxm = parser.load_hdxm('SecB_tetramer')
@@ -39,4 +30,14 @@ def test_load_from_yaml(self):
         assert isinstance(hdxm_set, HDXMeasurementSet)
         assert hdxm_set.names == list(data_dict.keys())
 
+    def test_load_job_parser(self):
+        fit_output_dir = input_dir / 'fit_result_output_1'
+        if fit_output_dir.exists():
+           shutil.rmtree(fit_output_dir, ignore_errors=True)
+
+        job_spec = yaml.safe_load((input_dir / 'jobfile.yaml').read_text())
+        parser = JobParser(job_spec, cwd=input_dir)
+        parser.execute()
 
+        assert fit_output_dir.exists()
+        shutil.rmtree(fit_output_dir, ignore_errors=True)

tests/test_data/input/jobfile.yaml

@@ -2,7 +2,7 @@ steps:
   - task: load_hdxm_set
     name: load_data
     state_file: data_states.yaml
-  - task: estimate_rates  # todo allow specification of method etc
+  - task: estimate_rates
     name: rates
     hdxm_set: $(load_data.out)
   - task: create_guess