Merge pull request #98 from LSYS/juancq

* From #73 by @juancq.

* Warn about duplicated `varlabel` (closes #76, closes #81).

* Add test that above warning works.

* Add known issues about duplicated `varlabel` (closes #76, closes #81) and PyCharm (closes #80).

Author: LSYS

README.md

@@ -308,6 +308,8 @@ More fined-grained control for base plot options (eg font sizes, marker colors)
 * Left-flushing of annotations relies on the `monospace` font.
 * Plot may give strange behavior for few rows of data (six rows or fewer. [see this issue](https://github.com/LSYS/forestplot/issues/52))
 * Plot can get cluttered with too many variables/rows (~30 onwards) 
+* Not tested with PyCharm (#80).
+* Duplicated `varlabel` may lead to unexpected results (see #76, #81). `mplot` for grouped models could be useful for such cases (see #59, WIP).
 <p align="right">(<a href="#top">back to top</a>)</p>
 
 <!----------------- BACKGROUND AND ADDITIONAL RESOURCES ----------------->

examples/subplot_example.py

@@ -0,0 +1,41 @@
+import matplotlib.gridspec as gridspec
+import matplotlib.pyplot as plt
+import numpy as np
+import pandas as pd
+
+import forestplot as fp
+
+# fill in
+df = pd.read_csv('test.csv')
+
+headers = ['Header 1','Header 2','Header 3','Header 4','Header 5']
+header_short = ['h1', 'h2', 'h3', 'h4', 'h5']
+
+fig, axarr = plt.subplots(2, 3, figsize=(20,20), sharey=True)
+fig.tight_layout(h_pad=2)
+
+k = 0
+for i in range(2):
+    for j in range(3):
+      col = header_short[k]
+      ax = fp.forestplot(df, ax=axarr[i,j], estimate=col+'1', ll=col+'2', hl=col+'3', 
+                         varlabel='label', 
+                  ci_report=False,  
+                  **{"marker": "D",  # set maker symbol as diamond
+                     "markersize": 35,  # adjust marker size
+                     "xlinestyle": (0, (10, 5)),  # long dash for x-reference line 
+                     "xlinecolor": ".1",  # gray color for x-reference line
+                     "xtick_size": 12,  # adjust x-ticker fontsize
+                     "fontsize": 14,
+                     }  )
+      if j > 0:
+         ax.axes.get_yaxis().set_visible(False)
+      ax.set_xlim(-.09, .09)
+      ax.set_title(headers[k])#, loc='left')
+      k += 1
+      if k >= 5:
+         break
+
+axarr[-1,-1].axis('off')
+
+plt.savefig('test.png', bbox_inches='tight', dpi=300)

forestplot/__init__.py

@@ -1,5 +1,5 @@
 """State version and import user-facing functions."""
-VERSION = (0, 3, 2)
+VERSION = (0, 3, 3)
 
 __version__ = ".".join(map(str, VERSION))
 

forestplot/arg_validators.py

@@ -221,6 +221,11 @@ def check_data(
                 "Duplicates found in variable labels ('varlabel') and group labels ('groupvar'). Formatting of y-axis labels may lead to unexpected errors."
             )
 
+    if len(dataframe) != dataframe[varlabel].dropna().nunique():
+        warnings.warn(
+            "Duplicates found in variable labels ('varlabel'). Plot may have errors."
+        )
+
     return dataframe
 
 

forestplot/plot.py

@@ -78,6 +78,7 @@ def forestplot(
     return_df: bool = False,
     preprocess: bool = True,
     table: bool = False,
+    ax: Optional[Axes] = None,
     **kwargs: Any,
 ) -> Axes:
     """
@@ -218,6 +219,7 @@ def forestplot(
         yticker2=yticker2,
         color_alt_rows=color_alt_rows,
         table=table,
+        ax=ax,
         **kwargs,
     )
     return (_local_df, ax) if return_df else ax
@@ -346,6 +348,7 @@ def _make_forestplot(
     yticker2: Optional[str],
     color_alt_rows: bool,
     figsize: Union[Tuple, List],
+    ax: Axes,
     despine: bool = True,
     table: bool = False,
     **kwargs: Any,
@@ -357,7 +360,8 @@ def _make_forestplot(
     -------
             Matplotlib Axes object.
     """
-    _, ax = plt.subplots(figsize=figsize, facecolor="white")
+    if not ax:
+        _, ax = plt.subplots(figsize=figsize, facecolor="white")
     ax = draw_ci(
         dataframe=dataframe,
         estimate=estimate,

setup.py

@@ -10,7 +10,7 @@
 install_requires = ["pandas", "numpy", "matplotlib", "matplotlib-inline<=0.1.3"]
 setup(
     name="forestplot",
-    version="0.3.2",
+    version="0.3.3",
     license="MIT",
     author="Lucas Shen",
     author_email="lucas@lucasshen.com",

tests/test_arg_validators.py

@@ -105,7 +105,7 @@ def test_check_data():
     check_data(
         dataframe=_df,
         estimate="estimate",
-        varlabel="moerror",
+        varlabel="estimate",
         annote=["ci_range"],
     )
 
@@ -147,7 +147,7 @@ def test_check_data():
     check_data(
         dataframe=_df,
         estimate="estimate",
-        varlabel="moerror",
+        varlabel="estimate",
         rightannote=["ci_range"],
     )
 
@@ -178,6 +178,20 @@ def test_check_data():
         == "Annotation headers are provided but no columns provided ('annote')."
     )
 
+    ##########################################################################
+    ## Check that warning for duplicated label works
+    ##########################################################################
+    duplicated_string = ["a", "a", "c"]
+    numeric = [-1, 2, 3.0]
+
+    # Assert that assertion for numeric type for estimate works
+    _df = pd.DataFrame({"varlabel": duplicated_string, "estimate": numeric})
+    with pytest.warns(UserWarning) as user_warning:
+        check_data(dataframe=_df, estimate="estimate", varlabel="varlabel")
+    assert "Duplicates found in variable labels ('varlabel'). Plot may have errors." in str(
+        user_warning[0].message
+    )
+
 
 def test_check_iterables_samelen():
     thresholds = (0.01, 0.05, 0.1)