Adjust test data for high-precision run

Author: jeromekelleher

sc2ts/inference.py

@@ -421,22 +421,22 @@ def match_samples(
     #         rerun_batch.append(sample)
 
     rerun_batch = samples
-    precision = 12
+    precision = 6
     logger.info(f"Rerunning batch of {len(rerun_batch)} at p={precision}")
     match_tsinfer(
         samples=rerun_batch,
         ts=base_ts,
         num_mismatches=num_mismatches,
-        precision=12,
+        precision=precision,
         num_threads=num_threads,
         show_progress=show_progress,
     )
-    # for sample in samples_to_rerun:
-    #     hmm_cost = sample.get_hmm_cost(num_mismatches)
-    #     # print(f"Final HMM pass:{sample.strain} hmm_cost={hmm_cost} path={sample.path}")
-    #     logger.debug(
-    #         f"Final HMM pass:{sample.strain} hmm_cost={hmm_cost} path={sample.path}"
-    #     )
+    for sample in rerun_batch:
+        hmm_cost = sample.get_hmm_cost(num_mismatches)
+        # print(f"Final HMM pass:{sample.strain} hmm_cost={hmm_cost} path={sample.path}")
+        logger.debug(
+            f"Final HMM pass:{sample.strain} hmm_cost={hmm_cost} path={sample.path}"
+        )
 
     # remaining_samples = samples
     # for cost, precision in [(0, 0), (1, 2)]: #, (2, 3)]:
@@ -463,7 +463,8 @@ def match_samples(
     #     remaining_samples = samples_to_rerun
 
     # Return in sorted order so that results are deterministic
-    return sorted(samples, key=lambda s: s.strain)
+    # return sorted(samples, key=lambda s: s.strain)
+    return samples
 
 
 def check_base_ts(ts):

tests/test_inference.py

@@ -612,17 +612,17 @@ def test_node_mutation_counts(self, fx_ts_map, date):
             "2020-01-30": {"nodes": 22, "mutations": 19},
             "2020-01-31": {"nodes": 23, "mutations": 21},
             "2020-02-01": {"nodes": 28, "mutations": 27},
-            "2020-02-02": {"nodes": 34, "mutations": 36},
-            "2020-02-03": {"nodes": 37, "mutations": 42},
-            "2020-02-04": {"nodes": 42, "mutations": 48},
-            "2020-02-05": {"nodes": 43, "mutations": 48},
-            "2020-02-06": {"nodes": 49, "mutations": 51},
-            "2020-02-07": {"nodes": 51, "mutations": 57},
-            "2020-02-08": {"nodes": 57, "mutations": 58},
-            "2020-02-09": {"nodes": 59, "mutations": 61},
-            "2020-02-10": {"nodes": 60, "mutations": 65},
-            "2020-02-11": {"nodes": 62, "mutations": 66},
-            "2020-02-13": {"nodes": 66, "mutations": 68},
+            "2020-02-02": {"nodes": 33, "mutations": 36},
+            "2020-02-03": {"nodes": 36, "mutations": 42},
+            "2020-02-04": {"nodes": 41, "mutations": 48},
+            "2020-02-05": {"nodes": 42, "mutations": 48},
+            "2020-02-06": {"nodes": 48, "mutations": 51},
+            "2020-02-07": {"nodes": 50, "mutations": 57},
+            "2020-02-08": {"nodes": 56, "mutations": 58},
+            "2020-02-09": {"nodes": 58, "mutations": 61},
+            "2020-02-10": {"nodes": 59, "mutations": 65},
+            "2020-02-11": {"nodes": 61, "mutations": 66},
+            "2020-02-13": {"nodes": 65, "mutations": 68},
         }
         assert ts.num_nodes == expected[date]["nodes"]
         assert ts.num_mutations == expected[date]["mutations"]
@@ -634,10 +634,10 @@ def test_node_mutation_counts(self, fx_ts_map, date):
             (7, "SRR11397729", 5),
             (13, "SRR11597132", 10),
             (16, "SRR11597177", 10),
-            (42, "SRR11597156", 10),
-            (57, "SRR11597216", 1),
-            (60, "SRR11597207", 41),
-            (62, "ERR4205570", 58),
+            (41, "SRR11597156", 10),
+            (56, "SRR11597216", 1),
+            (59, "SRR11597207", 40),
+            (61, "ERR4205570", 57),
         ],
     )
     def test_exact_matches(self, fx_ts_map, node, strain, parent):