Merge pull request #162 from jmccreight/bug_model_graph

fix and fix up model graphs

autotest/test_nhm_self_drive.py

@@ -1,9 +1,8 @@
 import pathlib as pl
 
-# import pytest
-import pywatershed as pws
 import xarray as xr
 
+import pywatershed as pws
 
 n_time_steps = 50
 

pywatershed/analysis/model_graph.py

@@ -1,6 +1,7 @@
 import pathlib as pl
 import tempfile
 import warnings
+from pprint import pprint
 
 from ..base.model import Model
 
@@ -34,6 +35,8 @@ def __init__(
         if not has_pydot:
             warnings.warn("pydot not available")
 
+        self.graph = None
+
         self.model = model
         self.show_params = show_params
         self.process_colors = process_colors
@@ -46,8 +49,6 @@ def __init__(
         self.node_spacing = node_spacing
         self.hide_variables = hide_variables
 
-        self.graph = None
-
         return
 
     def build_graph(self):
@@ -60,12 +61,13 @@ def build_graph(self):
             )
 
         # Solve the connections
-        files = []
-        connections = []
+        self.files = []
+        self.connections = []
         for process in self.model.process_order:
             frm_already = []
             for var, frm in self.model.process_input_from[process].items():
                 var_con = f":{var}"
+
                 if self.hide_variables:
                     var_con = ""
                     if frm in frm_already:
@@ -79,13 +81,14 @@ def build_graph(self):
                     color = self.default_edge_color
                     if self.process_colors:
                         color = self.process_colors[frm]
-                    connections += [
+                    self.connections += [
                         (f"{frm}{var_con}", f"{process}{var_con}", color)
                     ]
+
                 else:
                     file_name = frm.name
-                    files += [file_name]
-                    connections += [
+                    self.files += [file_name]
+                    self.connections += [
                         (
                             f"Files:{file_name.split('.')[0]}",
                             f"{process}{var_con}",
@@ -95,7 +98,7 @@ def build_graph(self):
 
         # Build the file node, reset the position
         self._current_pos = 0
-        self.file_node = self._file_node(files)
+        self.file_node = self._file_node(self.files)
 
         # build the graph
         self.graph = pydot.Dot(
@@ -109,20 +112,20 @@ def build_graph(self):
         for process in self.model.process_order:
             self.graph.add_node(self.process_nodes[process])
 
-        for con in connections:
+        for con in self.connections:
             self.graph.add_edge(pydot.Edge(con[0], con[1], color=con[2]))
 
         return
 
-    def SVG(self, verbose: bool = False):
+    def SVG(self, verbose: bool = False, dpi=45):
         """Display an SVG in jupyter notebook (via tempfile)."""
 
         if not has_ipython:
             warnings.warn("IPython is not available")
         tmp_file = pl.Path(tempfile.NamedTemporaryFile().name)
         if self.graph is None:
             self.build_graph()
-        self.graph.write_svg(tmp_file)
+        self.graph.write_svg(tmp_file, prog=["dot", f"-Gdpi={dpi}"])
         if verbose:
             print(f"Displaying SVG written to temp file: {tmp_file}")
 
@@ -183,7 +186,7 @@ def _process_node(self, process, show_params: bool = False):
                 if vv in mass_budget_vars:
                     border_color_str = 'border="1" COLOR="BLUE"'
                 label += f"    <TR>\n"
-                label += f'        <TD COLSPAN="4" BGCOLOR="{category_colors[varset_name]}" {border_color_str} PORT="{vv}" ><FONT POINT-SIZE="9.0">{vv}</FONT></TD>\n'
+                label += f'        <TD COLSPAN="4" BGCOLOR="{category_colors[varset_name]}" {border_color_str} PORT="{vv}"><FONT POINT-SIZE="9.0">{vv}</FONT></TD>\n'
                 label += f"    </TR>\n"
 
         label += f"</TABLE>>\n"

pywatershed/analysis/process_plot.py

@@ -73,8 +73,9 @@ def __init__(
         # HRU one-time setups
         self.hru_gdf = (
             gpd.read_file(self.hru_shapefile)
-            # .drop("nhm_id", axis=1)
-            .rename(columns={"nhru_v1_1": "nhm_id"}).set_index("nhm_id")
+            .drop("nhm_id", axis=1)
+            .rename(columns={"nhru_v1_1": "nhm_id"})
+            .set_index("nhm_id")
         )
 
         # segment one-time setup
@@ -94,15 +95,13 @@ def __init__(
         return
 
     def plot(self, var_name: str, process: StorageUnit, cmap: str = None):
-        var_dims = list(
-            meta.get_vars(var_name)[var_name]["dimensions"].values()
-        )
+        var_dims = list(meta.get_vars(var_name)[var_name]["dims"])
         if "nsegment" in var_dims:
             if not cmap:
                 cmap = "cool"
-            return self.plot_seg_var(self, var_name, process, cmap)
+            return self.plot_seg_var(var_name, process, cmap)
         elif "nhru" in var_dims:
-            return self.plot_hru_var(self, var_name, process)
+            return self.plot_hru_var(var_name, process)
         else:
             raise ValueError()
 
@@ -197,16 +196,23 @@ def get_hru_var(self, var_name: str, model: Model):
         ).set_index("nhm_id")
         return data_df
 
-    def plot_hru(
+    def plot_hru_var(
         self,
         var_name: str,
-        model: Model = None,
+        process: StorageUnit,
         data: np.ndarray = None,
         data_units: str = None,
         nhm_id: np.ndarray = None,
     ):
         if data is None:
-            data_df = self.get_hru_var(var_name, model)
+            # data_df = self.get_hru_var(var_name, model)
+            data_df = pd.DataFrame(
+                {
+                    "nhm_id": process.control.params.parameters["nhm_id"],
+                    var_name: process[var_name],
+                }
+            ).set_index("nhm_id")
+
         else:
             if nhm_id is None:
                 nhm_id = model.control.params.parameters["nhm_id"]

pywatershed/base/model.py

@@ -64,6 +64,7 @@ def __init__(
             vars = deepcopy(class_vars)
             c_inputs = inputs.pop(comp)
             _ = vars.pop(comp)
+
             self._inputs_from[comp] = {}
             # inputs_from_prev[comp] = {}
             # inputs
@@ -178,7 +179,6 @@ def _find_input_files(self):
                 load_n_time_batches=self._load_n_time_batches,
             )
         for process in self.process_order:
-            self.process_input_from[process] = {}
             for input, frm in self._inputs_from[process].items():
                 if not frm:
                     fname = file_inputs[input]._fname