Store low-energy coordinates in database

dimsim/_db/__init__.py

@@ -0,0 +1 @@
+"""A database tracking lowest-energy coordinates for a packed box of a given (physical property or compute) target."""

dimsim/_db/get.py

@@ -0,0 +1,17 @@
+import sqlite3
+
+
+def get_record(db_path: str, record_id: str) -> tuple[str, float]:
+    conn = sqlite3.connect(db_path)
+    cursor = conn.cursor()
+
+    cursor.execute("SELECT filename, energy FROM records WHERE id = ?", (record_id,))
+    row = cursor.fetchone()
+
+    conn.close()
+
+    if row is None:
+        raise KeyError(f"No record found with id={record_id!r}")
+
+    filename, energy = row
+    return filename, energy

dimsim/_db/insert.py

@@ -0,0 +1,11 @@
+import sqlite3
+
+
+def insert_record(db_path: str, record_id: str, filename: str, energy: float):
+    conn = sqlite3.connect(db_path)
+    cursor = conn.cursor()
+
+    cursor.execute("INSERT INTO records (id, filename, energy) VALUES (?, ?, ?)", (record_id, filename, energy))
+
+    conn.commit()
+    conn.close()

dimsim/_db/start.py

@@ -0,0 +1,19 @@
+import sqlite3
+
+
+def create_database(db_path: str):
+    conn = sqlite3.connect(db_path)
+    cursor = conn.cursor()
+
+    # not sure if we should be tracking file name (easy enough, brittle)
+    # or storing the full 3-D coordinates in the database
+    cursor.execute("""
+        CREATE TABLE IF NOT EXISTS records (
+            id TEXT PRIMARY KEY CHECK(length(id) = 16),
+            filename TEXT NOT NULL,
+            energy REAL NOT NULL
+        )
+    """)
+
+    conn.commit()
+    conn.close()

dimsim/compute/workflow.py

@@ -3,6 +3,7 @@
 
 import parsl
 
+from dimsim._db.start import create_database
 from dimsim.compute.apps import (
     minimize_energy,
     prepare_openmm_system,
@@ -24,6 +25,8 @@ def __init__(self, base_dir, parsl_config):
 
         parsl.load(parsl_config)
 
+        self.low_energy_database = create_database(str(pathlib.Path(self.base_dir) / "minima.db"))
+
     def _submit_compute(
         self,
         compute_config: BaseComputeConfig,
@@ -50,15 +53,21 @@ def _submit_compute(
         # trajectory_dcd = File(f"{job_dir}/trajectory.dcd")
 
         # 1. pack from compute config
+        # TODO: Short-circuit if a packed topolgoy already exists
         pack_future = prepare_packed_topology(compute_config, job_dir)
 
         # 2. set up openmm system
+        # TODO: Also short-circuit here
         setup_future = prepare_openmm_system(pack_future, job_dir)
 
         # 3. (for now ...) get minimized energy
+        # TODO: Short-circuit here too
+        # TODO: Store the resulting energy into the database, as a starting point
         minimize_future = minimize_energy(setup_future, job_dir)
 
         # 4. run equilibration step
+        # TODO: Short-circuit here too
+        # TODO: Store resulting energy into the database
         equilibration_future = run_equilibration(
             compute_config=compute_config,
             equilibration_config=None,
@@ -67,6 +76,7 @@ def _submit_compute(
         )
 
         # 5. run "production" step
+        # TODO: Store resulting energy into the database
         production_future = run_production(
             compute_config=compute_config,
             production_config=None,