adding labels to parallelDG

melmasri · Nov 9, 2022 · 9290cf4 · 9290cf4
1 parent 9538939
commit 9290cf4
Show file tree

Hide file tree

Showing 5 changed files with 21 additions and 7 deletions.
diff --git a/Dockerfile b/Dockerfile
@@ -5,7 +5,7 @@ WORKDIR /git
 RUN git clone https://github.com/melmasri/parallelDG.git
 WORKDIR /git/parallelDG
 RUN git fetch --all --tag # This is not triggered if the version is changed. It should be.
-RUN git checkout v0.8 -b latest
+RUN git checkout v0.9.2 -b latest
 RUN pip install -r requirements.txt
 ENV PYTHONPATH /git:/git/parallelDG:/git/parallelDG/bin
 ENV PATH /git/parallelDG/bin:$PATH

diff --git a/bin/parallelDG_loglinear_sample b/bin/parallelDG_loglinear_sample
@@ -1,7 +1,7 @@
 #!/usr/bin/env python
 
 """
-Generate a Markov chain from particle Gibbs for different
+Generate a Markov chain from Metropolis-Hastings trajectories
 parameter settings.
 """
 

diff --git a/parallelDG/auxiliary_functions.py b/parallelDG/auxiliary_functions.py
@@ -379,6 +379,7 @@ def plot_graph_traj_statistics(graph_traj, write_to_file=False):
 
 def write_traj_to_file(graph_trajectory,
                        dirt,
+                       labels,
                        output_filename=None,
                        output_format=None):
     date = datetime.datetime.today().strftime('%Y%m%d%H%m%S')
@@ -387,7 +388,7 @@ def write_traj_to_file(graph_trajectory,
     if output_format == 'benchpress':
         filename = dirt + "/"+output_filename
         filename_sub = dirt+"/"+output_filename[:-4]+'_subindex.csv'
-        df = graph_trajectory.graph_diff_trajectory_df()
+        df = graph_trajectory.graph_diff_trajectory_df(labels=labels)
         _cols_main = ['added', 'index', 'removed', 'score']
         _cols_sub = ['added_sub', 'subindex', 'removed_sub']
         df2 = df[_cols_main].dropna(axis=0)

diff --git a/parallelDG/graph/trajectory.py b/parallelDG/graph/trajectory.py
@@ -200,13 +200,18 @@ def default(o):
             with open(filename, 'w') as outfile:
                 json.dump(self.to_json(optional=optional), outfile, default=default)
 
-    def graph_diff_trajectory_df(self, subindex=True):
+    def graph_diff_trajectory_df(self, subindex=True, labels=None):
+
+        if labels is None:
+            p = self.init_graph.number_of_nodes()
+            labels = np.array(list(range(p)))
+
 
         def list_to_string(edge_list):
             s = "["
             for i, e in enumerate(edge_list):
                 e = sorted(e)
-                s += str(e[0]) + "-" + str(e[1])
+                s += str(labels[e[0]]) + "-" + str(labels[e[1]])
                 if i != len(edge_list)-1:
                     s += ";"
             return s + "]"

diff --git a/parallelDG/mh_parallel.py b/parallelDG/mh_parallel.py
@@ -494,6 +494,7 @@ def trajectory_to_file(n_samples,
                        reset_cache=True,
                        output_directory=".",
                        reseed=False,
+                       labels=None,
                        **args):
     """ Writes the trajectory of graphs generated by particle Gibbs to file.
 
@@ -521,10 +522,11 @@ def trajectory_to_file(n_samples,
     if not output_filename:
         output_filename = 'graph_traj_'+str(seed)+'.csv'
     output_format = args.get("output_format", None)
-    aux.write_traj_to_file(graph_trajectory,
+    aux.write_traj_to_file(graph_trajectory=graph_trajectory,
                            dirt=output_directory,
                            output_filename=output_filename,
-                           output_format=output_format
+                           output_format=output_format,
+                           labels=labels
     )
     return graph_trajectory
 
@@ -545,6 +547,8 @@ def sample_trajectories_ggm_to_file(dataframe,
     sd = seqdist.GGMJTPosterior()
     sd_graph = get_prior(graph_prior)
     graph_trajectories = []
+    node_labels = np.array(dataframe.columns.get_level_values(0))
+
     for _ in range(reps):
         for T in n_samples:
             for r in randomize:
@@ -557,6 +561,7 @@ def sample_trajectories_ggm_to_file(dataframe,
                                                           seqdist_graph=sd_graph,
                                                           reset_cache=reset_cache,
                                                           output_directory=output_directory,
+                                                          labels=node_labels,
                                                           **args)
                     graph_trajectories.append(graph_trajectory)
     return graph_trajectories
@@ -654,6 +659,8 @@ def sample_trajectories_loglin_to_file(dataframe,
     dt = dataframe.to_numpy()
     sd_graph = get_prior(graph_prior)
     graph_trajectories = []
+    node_labels = np.array(dataframe.columns.get_level_values(0))
+
     for _ in range(reps):
         for T in n_samples:
             for r in randomize:
@@ -666,6 +673,7 @@ def sample_trajectories_loglin_to_file(dataframe,
                                                           seqdist_graph=sd_graph,
                                                           reset_cache=reset_cache,
                                                           output_directory=output_directory,
+                                                          labels=node_labels,
                                                           **args)
                     graph_trajectories.append(graph_trajectory)
     return graph_trajectories