Merge pull request #361 from widdowquinn/pr_358

Issue #357: Fix genome labels in matrix output

.flake8

@@ -1,5 +1,5 @@
 [flake8]
-ignore = E203, E231, E266, E501, W503, F403, F401
+ignore = E203, E231, E266, E501, W503, F403, F401, E731
 max-line-length = 88
 max-complexity = 18
-select = B,C,E,F,W,T4,B9
+select = B,C,E,F,W,T4,B9

pyani/pyani_report.py

@@ -189,27 +189,25 @@ def write_dbtable(
     dfm: pd.DataFrame,
     path: Path,
     formats: Sequence[str] = ("tab",),
-    index: bool = False,
-    show_index: bool = False,
+    show_index: bool = True,
     colour_num: bool = False,
 ) -> None:
     """Write database result table to output file in named format.
 
     :param dfm:  pd.Dataframe
     :param path:  Path to output file
     :param formats:  tuple of str, output file formats
-    :param index:  Boolean
-    :param show_index:  Boolean
+    :param show_index:  output row and column labels
     :param colour_num:  use colours for values in HTML output
 
     colours are used for identity/coverage tables
     """
     formatdict = {
-        "tab": (dfm.to_csv, {"sep": "\t", "index": False}, ".tab"),
+        "tab": (dfm.to_csv, {"sep": "\t", "index": show_index}, ".tab"),
         "excel": (dfm.to_excel, {"index": show_index}, ".xlsx"),
         "html": (
             write_styled_html,
-            {"dfm": dfm, "index": index, "colour_num": colour_num},
+            {"dfm": dfm, "index": show_index, "colour_num": colour_num},
             ".html",
         ),
         "stdout": (write_to_stdout, {"dfm": dfm, "show_index": show_index}, ""),

pyani/pyani_tools.py

@@ -303,8 +303,13 @@ def label_results_matrix(matrix: pd.DataFrame, labels: Dict) -> pd.DataFrame:
     Applies the labels from the dictionary to the dataframe in
     matrix, and returns the result.
     """
-    matrix.columns = [f"{labels.get(_, _)}:{_}" for _ in matrix.columns]
-    matrix.index = [f"{labels.get(_, _)}:{_}" for _ in matrix.index]
+    # The dictionary uses string keys!
+    # Create a label function that produces <label>:<genome_id>
+    # when a label is available; and just Genome_id:<genome_id> when no
+    # label exists
+    label = lambda gen_id: f"{labels.get(str(gen_id), 'Genome_id')}:{gen_id}"
+    matrix.columns = [label(_) for _ in matrix.columns]
+    matrix.index = [label(_) for _ in matrix.index]
     return matrix
 
 

pyani/scripts/parsers/report_parser.py

@@ -120,11 +120,19 @@ def build(
         default=None,
         help="Report matrices of results for comma separated list of runs",
     )
+    parser.add_argument(
+        "--no_matrix_labels",
+        action="store_true",
+        dest="no_matrix_labels",
+        default=False,
+        help="Turn off row/column labels in output matrix files",
+    )
     parser.add_argument(
         "--formats",
         dest="formats",
         action="store",
         default=None,
+        choices=["html", "excel", "stdout"],
         help="Output formats (in addition to .tab)",
     )
     parser.set_defaults(func=subcommands.subcmd_report)

pyani/scripts/subcommands/subcmd_report.py

@@ -155,7 +155,10 @@ def subcmd_report(args: Namespace) -> int:
             "genome class",
         ]
         report(
-            args, session, formats, ReportParams("runs_genomes", statement, headers),
+            args,
+            session,
+            formats,
+            ReportParams("runs_genomes", statement, headers),
         )
 
     # Report table of all runs in which a genome is involved
@@ -194,7 +197,10 @@ def subcmd_report(args: Namespace) -> int:
             "date run",
         ]
         report(
-            args, session, formats, ReportParams("genomes_runs", statement, headers),
+            args,
+            session,
+            formats,
+            ReportParams("genomes_runs", statement, headers),
         )
 
     # Report table of comparison results for the indicated runs
@@ -259,6 +265,7 @@ def subcmd_report(args: Namespace) -> int:
     # JSON, we don't bother with a helper function like report(), and write out
     # our matrices directly, here
     if args.run_matrices:
+        show_index = not args.no_matrix_labels
         for run_id in [run_id.strip() for run_id in args.run_matrices.split(",")]:
             logger.debug("Extracting matrices for run %s", run_id)
             run = session.query(Run).filter(Run.run_id == run_id).first()
@@ -286,14 +293,19 @@ def subcmd_report(args: Namespace) -> int:
                         )
                     ),
                     formats,
-                    show_index=True,
+                    show_index=show_index,
                     **matdata.graphic_args,
                 )
 
     return 0
 
 
-def report(args: Namespace, session, formats: List[str], params: ReportParams,) -> None:
+def report(
+    args: Namespace,
+    session,
+    formats: List[str],
+    params: ReportParams,
+) -> None:
     """Write tabular report of pyani runs from database.
 
     :param args:  Namespace of command-line arguments

tests/test_subcmd_05_report.py

@@ -81,6 +81,7 @@ def setUp(self):
                 show_genomes_runs=False,
                 run_results=False,
                 run_matrices=False,
+                no_matrix_labels=False,
                 force=True,
                 formats="html,excel,stdout",
             ),
@@ -93,6 +94,7 @@ def setUp(self):
                 show_genomes_runs=False,
                 run_results=False,
                 run_matrices=False,
+                no_matrix_labels=False,
                 force=True,
                 formats="html,excel,stdout",
             ),
@@ -105,6 +107,7 @@ def setUp(self):
                 show_genomes_runs=False,
                 run_results=False,
                 run_matrices=False,
+                no_matrix_labels=False,
                 force=True,
                 formats="html,excel,stdout",
             ),
@@ -117,6 +120,7 @@ def setUp(self):
                 show_genomes_runs=True,
                 run_results=False,
                 run_matrices=False,
+                no_matrix_labels=False,
                 force=True,
                 formats="html,excel,stdout",
             ),
@@ -129,6 +133,7 @@ def setUp(self):
                 show_genomes_runs=False,
                 run_results="1",
                 run_matrices=False,
+                no_matrix_labels=False,
                 force=True,
                 formats="html,excel,stdout",
             ),
@@ -141,6 +146,20 @@ def setUp(self):
                 show_genomes_runs=False,
                 run_results=False,
                 run_matrices="1",
+                no_matrix_labels=False,
+                force=True,
+                formats="html,excel",
+            ),
+            "no_matrix_labels": Namespace(
+                outdir=self.outdir,
+                dbpath=self.dbpath,
+                show_runs=False,
+                show_genomes=True,
+                show_runs_genomes=False,
+                show_genomes_runs=False,
+                run_results=False,
+                run_matrices="1",
+                no_matrix_labels=True,
                 force=True,
                 formats="html,excel",
             ),
@@ -169,3 +188,7 @@ def test_results(self):
     def test_matrices(self):
         """Test reporting of run matrices in the database."""
         subcommands.subcmd_report(self.argsdict["run_matrices"])
+
+    def test_no_matrix_labels(self):
+        """Test row and column labeling of run matrices."""
+        subcommands.subcmd_report(self.argsdict["no_matrix_labels"])