Allow no drawing of CI (#61)

* Allow no drawing of CI (#58)

* Update CI.yml

* Update nb.yml

* Update options (#58)

* Increase images

* Bump to v0.3.0 (#58)

.github/workflows/CI.yml

@@ -2,7 +2,7 @@ name: CI
 
 on:
   push:
-    branches: [ "main", "docs", "patch", "feature" ]
+    branches: [ "main", "docs", "patch", "feature", "nodrawci" ]
   pull_request:
     branches: [ "main" ]
 

.github/workflows/nb.yml

@@ -2,7 +2,7 @@ name: Notebooks
 
 on:
   push:
-    branches: [ "main", "patch", "feature" ]
+    branches: [ "main", "patch", "feature", "nodrawci" ]
   pull_request:
     branches: [ "main" ]
 
@@ -13,7 +13,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        python-version: ['3.9']
+        python-version: ['3.10']
     steps:
     - uses: actions/checkout@v3
     - name: Set up Python ${{ matrix.python-version }}

README.md

@@ -98,13 +98,12 @@ df.head(3)
 
   | Column    | Description                                     | Required  |
   |:----------|:------------------------------------------------|:----------|
-  | `var`     | Variable field                                  |           |
+  | `var`     | Variable label                                  | ✓   |
   | `r`       | Correlation coefficients (estimates to plot)    | ✓   |
-  | `moerror` | Conf. int.'s *margin of error*.                 |           |
   | `label`   | Variable labels                                 | ✓   |
   | `group`   | Variable grouping labels                        |           |
-  | `ll`      | Conf. int. *lower limits*                       | ✓  |
-  | `hl`      | Containing the conf. int. *higher limits*       | ✓  |
+  | `ll`      | Conf. int. *lower limits*                       |           |
+  | `hl`      | Containing the conf. int. *higher limits*       |           |
   | `n`       | Sample size                                     |           |
   | `power`   | Statistical power                               |           |
   | `p-val`   | P-value                                         |           |
@@ -122,7 +121,7 @@ fp.forestplot(df,  # the dataframe with results data
               xlabel="Pearson correlation",  # x-label title
               )
 ```
-<p align="left"><img width="55%" src="https://raw.githubusercontent.com/LSYS/forestplot/main/docs/images/vanilla.png"></p>
+<p align="left"><img width="75%" src="https://raw.githubusercontent.com/LSYS/forestplot/main/docs/images/vanilla.png"></p>
 
 Save the plot
 ```python
@@ -202,7 +201,7 @@ fp.forestplot(df,  # the dataframe with results data
               **{'fontfamily': 'sans-serif'}  # revert to sans-serif                              
               )
 ```               
-<p align="left"><img width="40%" src="https://raw.githubusercontent.com/LSYS/forestplot/main/docs/images/vcoefplot.png"></p>
+<p align="left"><img width="50%" src="https://raw.githubusercontent.com/LSYS/forestplot/main/docs/images/vcoefplot.png"></p>
 
 5. Example with more customizations
 ```python
@@ -232,7 +231,7 @@ fp.forestplot(df,  # the dataframe with results data
                 }  
               )
 ```
-<p align="left"><img width="80%" src="https://raw.githubusercontent.com/LSYS/forestplot/main/docs/images/main.png"></p>
+<p align="left"><img width="100%" src="https://raw.githubusercontent.com/LSYS/forestplot/main/docs/images/main.png"></p>
 
 <details><summary><i>Annotations arguments allowed include:</i></summary>
 
@@ -259,8 +258,6 @@ fp.forestplot(df,  # the dataframe with results data
 <p align="right">(<a href="#top">back to top</a>)</p>
 
 
-
-
 <!------------------- GALLERY AND API OPTIONS ------------------->
 ## Gallery and API Options[![](https://raw.githubusercontent.com/LSYS/forestplot/main/docs/images/pin.svg)](#gallery-and-api-options)
 
@@ -276,8 +273,8 @@ More fined-grained control for base plot options (eg font sizes, marker colors)
 | `dataframe`           | Pandas dataframe where rows are variables (or studies for meta-analyses) and columns include estimated effect sizes, labels, and confidence intervals, etc. | &check; |
 | `estimate`            | Name of column in `dataframe` containing the *estimates*.                                                                                                   | &check; |
 | `varlabel`            | Name of column in `dataframe` containing the *variable labels* (study labels if meta-analyses).                                                             | &check; |
-| `ll`                  | Name of column in `dataframe` containing the conf. int. *lower limits*.                                                                                     | &check; |
-| `hl`                  | Name of column in `dataframe` containing the conf. int. *higher limits*.                                                                                    | &check; |
+| `ll`                  | Name of column in `dataframe` containing the conf. int. *lower limits*.                                                                                     |  |
+| `hl`                  | Name of column in `dataframe` containing the conf. int. *higher limits*.                                                                                    |  |
 | `logscale`            | If True, make the x-axis log scale. Default is False.                                                                                                     |  |
 | `capitalize`          | How to capitalize strings. Default is None. One of "capitalize", "title", "lower", "upper", "swapcase".                                                      | |
 | `form_ci_report`      | If True (default), report the estimates and confidence interval beside the variable labels.                                                                 |          |

forestplot/__init__.py

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

forestplot/arg_validators.py

@@ -14,6 +14,7 @@ def check_data(
     group_order: Optional[Sequence] = None,
     ll: Optional[str] = None,
     hl: Optional[str] = None,
+    form_ci_report: bool = None,
     annote: Optional[Union[Sequence[str], None]] = None,
     annoteheaders: Optional[Union[Sequence[str], None]] = None,
     rightannote: Optional[Union[Sequence[str], None]] = None,
@@ -43,6 +44,8 @@ def check_data(
             Name of column containing the lower limit of the confidence intervals.
     hl (str)
             Name of column containing the upper limit of the confidence intervals.
+    form_ci_report (bool)
+            If True, form the formatted confidence interval as a string.
     annote (list-like)
             List of columns to add as additional annotation in the plot.
     annoteheaders (list-like)
@@ -61,6 +64,24 @@ def check_data(
     -------
             pd.core.frame.DataFrame.
     """
+    ##########################################################################
+    ## Check that CI options (ll, hl, form_ci_report) are consistent
+    ##########################################################################
+    if ll is None:
+        try:
+            assert hl is None
+        except Exception:
+            raise TypeError("'ll' is None. 'hl' should also be None.")
+
+    if hl is None:
+        try:
+            assert ll is None
+        except Exception:
+            raise TypeError("'hl' is None. 'll' should also be None.")
+
+    if ll is None and form_ci_report:
+        warnings.warn("'ll' is None. 'form_ci_report' will be set to False.")
+
     ##########################################################################
     ## Check that numeric data are numeric
     ##########################################################################

forestplot/graph_utils.py

@@ -43,17 +43,18 @@ def draw_ci(
     -------
             Matplotlib Axes object.
     """
-    lw = kwargs.get("lw", 1.4)
-    linecolor = kwargs.get("linecolor", ".6")
-    ax.errorbar(
-        x=dataframe[estimate],
-        y=dataframe[yticklabel],
-        xerr=[dataframe[estimate] - dataframe[ll], dataframe[hl] - dataframe[estimate]],
-        ecolor=linecolor,
-        elinewidth=lw,
-        ls="none",
-        zorder=0,
-    )
+    if ll is not None:
+        lw = kwargs.get("lw", 1.4)
+        linecolor = kwargs.get("linecolor", ".6")
+        ax.errorbar(
+            x=dataframe[estimate],
+            y=dataframe[yticklabel],
+            xerr=[dataframe[estimate] - dataframe[ll], dataframe[hl] - dataframe[estimate]],
+            ecolor=linecolor,
+            elinewidth=lw,
+            ls="none",
+            zorder=0,
+        )
     if logscale:
         ax.set_xscale("log", base=10)
     return ax
@@ -532,6 +533,7 @@ def format_xlabel(xlabel: str, ax: Axes, **kwargs: Any) -> Axes:
 
 def format_xticks(
     dataframe: pd.core.frame.DataFrame,
+    estimate: str,
     ll: str,
     hl: str,
     xticks: Optional[Union[list, range]],
@@ -550,6 +552,9 @@ def format_xticks(
     dataframe (pandas.core.frame.DataFrame)
             Pandas DataFrame where rows are variables. Columns are variable name, estimates,
             margin of error, etc.
+    estimate (str)
+            Name of column containing the estimates (e.g. pearson correlation coefficient,
+            OR, regression estimates, etc.).
     ll (str)
             Name of column containing the lower limit of the confidence intervals.
             Optional
@@ -568,8 +573,12 @@ def format_xticks(
     nticks = kwargs.get("nticks", 5)
     xtick_size = kwargs.get("xtick_size", 10)
     xticklabels = kwargs.get("xticklabels", None)
-    xlowerlimit = dataframe[ll].min()
-    xupperlimit = dataframe[hl].max()
+    if ll is not None:
+        xlowerlimit = dataframe[ll].min()
+        xupperlimit = dataframe[hl].max()
+    else:
+        xlowerlimit = 1.1 * dataframe[estimate].min()
+        xupperlimit = 1.1 * dataframe[estimate].max()
     ax.set_xlim(xlowerlimit, xupperlimit)
     if xticks is not None:
         ax.set_xticks(xticks)
@@ -653,6 +662,7 @@ def draw_tablelines(
     pval: str,
     right_annoteheaders: Optional[Union[Sequence[str], None]],
     ax: Axes,
+    **kwargs: Any
 ) -> Axes:
     """
     Plot horizontal lines as table lines.
@@ -684,7 +694,7 @@ def draw_tablelines(
         [x0, x1], [nrows - 1.45, nrows - 1.45], color="0.5", linewidth=lower_lw, clip_on=False
     )
     if (right_annoteheaders is not None) or (pval is not None):
-        extrapad = 0.05
+        extrapad = kwargs.get("extrapad", 0.05)
         x0 = ax.get_xlim()[1] * (1 + extrapad)
         plt.plot(
             [x0, righttext_width],

forestplot/plot.py

@@ -100,7 +100,7 @@ def forestplot(
     form_ci_report (bool)
             If True, form the formatted confidence interval as a string.
     ci_report (bool)
-            If True, form the formatted confidence interval as a string.
+            If True, report the formatted confidence interval as a string.
     groupvar (str)
             Name of column containing group of variables.
     group_order (list-like)
@@ -170,8 +170,8 @@ def forestplot(
         rightannote=rightannote,
         right_annoteheaders=right_annoteheaders,
     )
-    if (ll is None) or (hl is None):
-        ll, hl = "ll", "hl"
+    if ll is None:
+        ci_report = False
     if ci_report is True:
         form_ci_report = True
     if preprocess:
@@ -371,7 +371,9 @@ def _make_forestplot(
     draw_est_markers(
         dataframe=dataframe, estimate=estimate, yticklabel=yticklabel, ax=ax, **kwargs
     )
-    format_xticks(dataframe=dataframe, ll=ll, hl=hl, xticks=xticks, ax=ax, **kwargs)
+    format_xticks(
+        dataframe=dataframe, estimate=estimate, ll=ll, hl=hl, xticks=xticks, ax=ax, **kwargs
+    )
     draw_ref_xline(
         ax=ax,
         dataframe=dataframe,

forestplot/text_utils.py

@@ -49,16 +49,21 @@ def form_est_ci(
     -------
             pd.core.frame.DataFrame with an additional formatted 'est_ci' column.
     """
-    for col in [estimate, ll, hl]:
+    if ll is None:
+        cols = [estimate]
+    else:
+        cols = [estimate, ll, hl]
+    for col in cols:
         dataframe = _right_justify_num(
             dataframe=dataframe, col=col, decimal_precision=decimal_precision
         )
     for ix, row in dataframe.iterrows():
         formatted_est = row[f"formatted_{estimate}"]
-        formatted_ll, formatted_hl = row[f"formatted_{ll}"], row[f"formatted_{hl}"]
-        formatted_ci = "".join([caps[0], formatted_ll, connector, formatted_hl, caps[1]])
-        dataframe.loc[ix, "ci_range"] = formatted_ci
-        dataframe.loc[ix, "est_ci"] = "".join([formatted_est, formatted_ci])
+        if ll is not None:
+            formatted_ll, formatted_hl = row[f"formatted_{ll}"], row[f"formatted_{hl}"]
+            formatted_ci = "".join([caps[0], formatted_ll, connector, formatted_hl, caps[1]])
+            dataframe.loc[ix, "ci_range"] = formatted_ci
+            dataframe.loc[ix, "est_ci"] = "".join([formatted_est, formatted_ci])
     return dataframe
 
 

setup.py

@@ -10,7 +10,7 @@
 install_requires = ["pandas", "numpy", "matplotlib", "matplotlib-inline<=0.1.3"]
 setup(
     name="forestplot",
-    version="0.2.2",
+    version="0.3.0",
     license="MIT",
     author="Lucas Shen",
     author_email="lucas@lucasshen.com",

tests/test_arg_validators.py

@@ -25,9 +25,9 @@ def test_check_data():
     check_data(dataframe=_df, estimate="estimate", varlabel="varlabel")
 
     # Assert that assertion for numeric type for ll works
-    _df = pd.DataFrame({"estimate": numeric, "ll": string})
+    _df = pd.DataFrame({"estimate": numeric, "ll": string, "hl": numeric})
     with pytest.raises(TypeError) as excinfo:
-        check_data(dataframe=_df, estimate="estimate", varlabel="estimate", ll="ll")
+        check_data(dataframe=_df, estimate="estimate", varlabel="estimate", ll="ll", hl="hl")
     assert str(excinfo.value) == "CI lowerlimit values should be float or int"
 
     # Assert that conversion for numeric ll stored as string works
@@ -41,9 +41,9 @@ def test_check_data():
     check_data(dataframe=_df, estimate="estimate", varlabel="estimate", ll="ll", hl="hl")
 
     # Assert that assertion for numeric type for hl works
-    _df = pd.DataFrame({"estimate": numeric, "hl": string})
+    _df = pd.DataFrame({"estimate": numeric, "ll": numeric, "hl": string})
     with pytest.raises(TypeError) as excinfo:
-        check_data(dataframe=_df, estimate="estimate", varlabel="estimate", hl="hl")
+        check_data(dataframe=_df, estimate="estimate", varlabel="estimate", ll="ll", hl="hl")
     assert str(excinfo.value) == "CI higherlimit values should be float or int"
 
     # Assert that conversion for numeric hl stored as string works
@@ -56,6 +56,17 @@ def test_check_data():
     )
     check_data(dataframe=_df, estimate="estimate", varlabel="estimate", ll="ll", hl="hl")
 
+    # Assert that check for CI options are consistent works
+    _df = pd.DataFrame({"estimate": numeric, "ll": string, "hl": string})
+    with pytest.raises(TypeError) as excinfo:
+        check_data(dataframe=_df, estimate="estimate", varlabel="estimate", ll=None, hl="hl")
+    assert str(excinfo.value) == "'ll' is None. 'hl' should also be None."
+
+    _df = pd.DataFrame({"estimate": numeric, "ll": string, "hl": string})
+    with pytest.raises(TypeError) as excinfo:
+        check_data(dataframe=_df, estimate="estimate", varlabel="estimate", ll="ll", hl=None)
+    assert str(excinfo.value) == "'hl' is None. 'll' should also be None."
+
     ##########################################################################
     ## Check annote
     ##########################################################################

tests/test_graph_utils.py

@@ -230,7 +230,7 @@ def test_format_xticks():
     )
     # No ticks set
     _, ax = plt.subplots()
-    ax = format_xticks(input_df, ll="ll", hl="hl", xticks=None, ax=ax)
+    ax = format_xticks(input_df, estimate="estimate", ll="ll", hl="hl", xticks=None, ax=ax)
     assert isinstance(ax, Axes)
     ax_xmin, ax_xmax = ax.get_xlim()
     data_xmin, data_xmax = input_df.ll.min(), input_df.hl.max()
@@ -239,7 +239,9 @@ def test_format_xticks():
 
     # Set xticks
     _, ax = plt.subplots()
-    ax = format_xticks(input_df, ll="ll", hl="hl", xticks=[1, 2, 3], ax=ax)
+    ax = format_xticks(
+        input_df, estimate="estimate", ll="ll", hl="hl", xticks=[1, 2, 3], ax=ax
+    )
     assert isinstance(ax, Axes)
     ax_xmin, ax_xmax = ax.get_xlim()
     data_xmin, data_xmax = input_df.ll.min(), input_df.hl.max()