diff --git a/flepimop/gempyor_pkg/setup.cfg b/flepimop/gempyor_pkg/setup.cfg
index 3de937cb7..ce7fd7f1b 100644
--- a/flepimop/gempyor_pkg/setup.cfg
+++ b/flepimop/gempyor_pkg/setup.cfg
@@ -38,6 +38,7 @@ install_requires =
 [options.entry_points]
 console_scripts =
     gempyor-outcomes = gempyor.simulate_outcome:simulate
+    flepimop = gempyor.cli:cli
     gempyor-seir = gempyor.simulate_seir:simulate
     gempyor-simulate = gempyor.simulate:simulate
 
diff --git a/flepimop/gempyor_pkg/src/gempyor/compartments.py b/flepimop/gempyor_pkg/src/gempyor/compartments.py
index 4ccb32e89..b75a0ab45 100644
--- a/flepimop/gempyor_pkg/src/gempyor/compartments.py
+++ b/flepimop/gempyor_pkg/src/gempyor/compartments.py
@@ -2,7 +2,7 @@
 import pandas as pd
 import pyarrow as pa
 import pyarrow.parquet as pq
-
+import click
 from .utils import config, Timer, as_list
 from . import file_paths
 from functools import reduce
@@ -248,10 +248,13 @@ def parse_single_transition(self, seir_config, single_transition_config, fake_co
 
         return rc
 
-    def toFile(self, compartments_file, transitions_file):
+    def toFile(self, compartments_file, transitions_file, write_parquet=False):
         out_df = self.compartments.copy()
-        pa_df = pa.Table.from_pandas(out_df, preserve_index=False)
-        pa.parquet.write_table(pa_df, compartments_file)
+        if write_parquet:
+            pa_df = pa.Table.from_pandas(out_df, preserve_index=False)
+            pa.parquet.write_table(pa_df, compartments_file)
+        else:
+            out_df.to_csv(compartments_file, index=False)
 
         out_df = self.transitions.copy()
         out_df["source"] = self.format_source(out_df["source"])
@@ -259,9 +262,11 @@ def toFile(self, compartments_file, transitions_file):
         out_df["rate"] = self.format_rate(out_df["rate"])
         out_df["proportional_to"] = self.format_proportional_to(out_df["proportional_to"])
         out_df["proportion_exponent"] = self.format_proportion_exponent(out_df["proportion_exponent"])
-        pa_df = pa.Table.from_pandas(out_df, preserve_index=False)
-        pa.parquet.write_table(pa_df, transitions_file)
-
+        if write_parquet:
+            pa_df = pa.Table.from_pandas(out_df, preserve_index=False)
+            pa.parquet.write_table(pa_df, transitions_file)
+        else:
+            out_df.to_csv(transitions_file, index=False)
         return
 
     def fromFile(self, compartments_file, transitions_file):
@@ -489,8 +494,8 @@ def parse_parameter_strings_to_numpy_arrays_v2(self, parameters, parameter_names
             # TODO: instead of searching for the next array, better to just use the parameter shape.
             if not isinstance(substituted_formulas[i], np.ndarray):
                 for k in range(len(substituted_formulas)): 
-                     if isinstance(substituted_formulas[k], np.ndarray):
-                         substituted_formulas[i] = substituted_formulas[i] * np.ones_like(substituted_formulas[k])
+                    if isinstance(substituted_formulas[k], np.ndarray):
+                        substituted_formulas[i] = substituted_formulas[i] * np.ones_like(substituted_formulas[k])
 
         return np.array(substituted_formulas)
 
@@ -643,3 +648,42 @@ def list_recursive_convert_to_string(thing):
     if type(thing) == list:
         return [list_recursive_convert_to_string(x) for x in thing]
     return str(thing)
+
+
+
+@click.group()
+def compartments():
+    pass
+
+# TODO: CLI arguments
+@compartments.command()
+def plot():
+    assert config["compartments"].exists() 
+    assert config["seir"].exists() 
+    comp = Compartments(seir_config=config["seir"], compartments_config=config["compartments"])
+    
+    # TODO: this should be a command like build compartments.
+    (
+        unique_strings,
+        transition_array,
+        proportion_array,
+        proportion_info,
+    ) = comp.get_transition_array()
+    
+    comp.plot(output_file="transition_graph", source_filters=[], destination_filters=[])
+
+    print("wrote file transition_graph")
+
+@compartments.command()
+def export():
+    assert config["compartments"].exists() 
+    assert config["seir"].exists() 
+    comp = Compartments(seir_config=config["seir"], compartments_config=config["compartments"])
+    (
+        unique_strings,
+        transition_array,
+        proportion_array,
+        proportion_info,
+    ) = comp.get_transition_array()
+    comp.toFile('compartments_file.csv', 'transitions_file.csv')
+    print("wrote files 'compartments_file.csv', 'transitions_file.csv' ")
\ No newline at end of file