added oracle method

lib/convergence_detection.ml

@@ -173,3 +173,18 @@ let plot_convergent_sites ?(plot_all_sites = true) ~alignment ~detection_results
     )
   ]
   |> Fn.flip Workflow.select ["plot.svg"]
+
+let recall_precision_curve table =
+  Workflow.shell ~descr:"recall_precision_curve" [
+    cmd "Rscript" [
+      file_dump (string Scripts.recall_precision_curve) ;
+      dep table ;
+      dest ;
+    ] ;
+  ]
+
+let%pworkflow oracle ~n_h0 ~n_ha =
+  "Sites\tOracle"
+  :: (List.init n_h0 ~f:(fun i -> sprintf "%d\t0" (i + 1)))
+  @ (List.init n_ha ~f:(fun i -> sprintf "%d\t1" (n_h0 + i + 1)))
+  |> Out_channel.write_lines [%dest]

lib/convergence_detection.mli

@@ -57,3 +57,12 @@ val plot_convergent_sites :
   tree:nhx pworkflow ->
   unit ->
   svg pworkflow
+
+val recall_precision_curve :
+  text_file pworkflow ->
+  svg pworkflow
+
+val oracle :
+  n_h0:int ->
+  n_ha:int ->
+  text_file pworkflow

lib/pipeline2.ml

@@ -116,3 +116,21 @@ let alignment_plot d =
     ~alignment:(amino_acid_alignment d)
     ~detection_results:(multinomial d)
     ()
+
+let oracle d =
+  let n_h0, n_ha =
+    match d with
+    | Bppseqgen_simulation { nb_sites ; hypothesis ; _ } -> (
+        match hypothesis with
+        | H0 _ -> nb_sites, 0
+        | HaPC _ | HaPCOC _ -> 0, nb_sites
+      )
+    | Convdet_simulation { n_h0 ; n_ha ; _ } -> n_h0, n_ha
+  in
+  Convergence_detection.oracle ~n_h0 ~n_ha
+
+let benchmark d =
+  Convergence_detection.merge_result_tables
+    ~multinomial:(multinomial d)
+    ~oracle:(oracle d)
+    ()

lib/scripts/merge_det_results.py

@@ -76,6 +76,8 @@ availableOptions.add_argument('--msd', type=str, action="append", nargs=2,
                    help="msd output name", default = None)
 availableOptions.add_argument('--fna_infos', type=str,
                    help="fna_infos output name", default = None)
+availableOptions.add_argument('--oracle', type=str,
+                   help="ORACLE", default = None)
 
 ##############
 
@@ -96,6 +98,7 @@ df_tdg09 = pd.DataFrame()
 df_multinomial = pd.DataFrame()
 df_msd = pd.DataFrame()
 df_fna_infos = pd.DataFrame()
+df_oracle = pd.DataFrame ()
 
 OutName = args.output
 if args.pcoc :
@@ -124,6 +127,10 @@ if args.diffsel_bis :
 if args.diffsel :
   df_diffsel = pd.read_csv(args.diffsel, sep="\t")
 
+if args.oracle:
+  df_oracle = pd.read_csv(args.oracle, sep="\t")
+  df_oracle.rename(columns={'Oracle' : 'Oracle'}, inplace=True)
+
 if args.identical_LG :
   df_identical_LG = pd.read_csv(args.identical_LG, sep="\t")
   df_identical_LG.rename(columns={'Identical': 'Identical_LG08'}, inplace=True)
@@ -165,13 +172,14 @@ if args.fna_infos :
   #df_fna_infos = df_fna_infos[['Sites','P_distance']]
 
 df_list = [df for df in [df_pcoc, df_pcoc_gamma, df_pcoc_C60,
-                        df_diffsel, df_diffsel_bis,
-                        df_identical_LG, df_identical_WAG,
-                        df_topological_LG, df_topological_WAG,
-                        df_tdg09,
-                        df_multinomial,
-                        df_msd,
-                        df_fna_infos
+                         df_diffsel, df_diffsel_bis,
+                         df_identical_LG, df_identical_WAG,
+                         df_topological_LG, df_topological_WAG,
+                         df_tdg09,
+                         df_multinomial,
+                         df_msd,
+                         df_fna_infos,
+                         df_oracle
                         ] if not df.empty ]
 
 df_list_len = [df.shape[0] for df in df_list]