go_utils.plot

  1import logging
  2import math
  3
  4import matplotlib.pyplot as plt
  5import pandas as pd
  6import seaborn as sns
  7
  8
  9def plot_freq_bar(df, protocol, column, title, plot_type="bar", log_scale=False):
 10    """
 11    Plots the frequency of different values of a column across the dataset
 12
 13    Parameters
 14    ----------
 15    df : pd.DataFrame
 16      The DataFrame containing the desired data.
 17    protocol : str
 18      The Protocol of the DataFrame (e.g. Mosquito Habitat Mapper or Land Cover)
 19    column : str
 20      The name of the column that the graph will plot.
 21    title : str
 22      The title of the graph
 23    plot_type : str, default="bar"
 24      The type of graph that will be used to visually compare the data.
 25    log_scale : bool, default=False
 26      If a log scale will be used for the data. If true, all the values will be in Log 10.
 27    """
 28    col_vals = pd.DataFrame()
 29    col_vals["Count"] = df.groupby(column).size()
 30    col_vals.reset_index(inplace=True)
 31    col_vals.columns = ["index", "Count"]
 32    if len(col_vals) <= 1:  # pragma: no cover
 33        logging.warning(
 34            f"There is only one value for this column: {col_vals['index'][0]}: {col_vals['Count'][0]}"
 35        )
 36    plt.figure(figsize=(10, 6))
 37    ylabel = "Frequency"
 38    title = f"{protocol} - {title}"
 39    if log_scale:
 40        col_vals["Count"] = pd.Series([math.log10(val) for val in col_vals["Count"]])
 41        ylabel += " (Log Scale)"
 42        title += " (Log Scale)"
 43    if plot_type == "line":  # pragma: no cover
 44        plt.plot(col_vals["index"], col_vals["Count"], color="lightblue")
 45    else:
 46        plt.bar(col_vals["index"], col_vals["Count"], color="lightblue")
 47
 48    plt.title(title)
 49    plt.xlim(left=-0.5)
 50    plt.xlabel(f"{column} Values")
 51    plt.ylabel(ylabel)
 52
 53
 54def multiple_bar_graph(df, protocol, cols, title, log_scale=False):
 55    """
 56    Plots the frequency of different values of a column across the dataset alongside eachother.
 57
 58    Parameters
 59    ----------
 60    df : pd.DataFrame
 61      The DataFrame containing the desired data.
 62    protocol : str
 63      The Protocol of the DataFrame (e.g. Mosquito Habitat Mapper or Land Cover)
 64    cols : list str
 65      The names of the columns that the graph will plot.
 66    title : str
 67      The title of the graph
 68    log_scale : bool, default=False
 69      If a log scale will be used for the data. If true, all the values will be in Log 10.
 70    """
 71
 72    def create_summary_df(cols):
 73        """
 74        data = df
 75        x = [0,6]
 76        hue = photoType: photocount, rejectedcount, etc
 77        """
 78        photo_summary = pd.DataFrame()
 79        for name in cols:
 80            photo_summary[name] = df[name].value_counts()
 81            # print(type(photo_summary[name]))
 82        photo_summary.sort_index(inplace=True)
 83        photo_summary = photo_summary.reset_index()
 84        # print(photo_summary)
 85
 86        category_counts = []
 87
 88        for name in cols:
 89            for j in range(len(photo_summary)):
 90                count = photo_summary[name][j]
 91                if log_scale:
 92                    count = math.log10(count)
 93                new_row = {
 94                    "index": photo_summary["index"][j],
 95                    "category": name,
 96                    "count": count,
 97                }
 98                category_counts.append(new_row)
 99        category_counts = pd.DataFrame(category_counts)
100        return category_counts
101
102    category_counts = create_summary_df(cols)
103    plt.figure(figsize=(10, 6))
104
105    title = f"{protocol} -- {title}"
106    ylabel = "Frequency"
107    if log_scale:
108        ylabel += " (Log Scale)"
109        title += " (Log Scale)"
110
111    ax = sns.barplot(
112        data=category_counts,
113        x="index",
114        y="count",
115        hue="category",
116        palette=[
117            "#377eb8",
118            "#ff7f00",
119            "#4daf4a",
120            "#f781bf",
121            "#a65628",
122            "#984ea3",
123            "#999999",
124            "#e41a1c",
125            "#dede00",
126        ],
127    )
128    ax.set_xlabel("Photo Count")
129    ax.set_ylabel(ylabel)
130
131    ax.set_title(title)
132    plt.legend(loc="upper right")
133
134
135def plot_int_distribution(df, col_name, title_name):
136    """
137    Plots the frequency of different integer values of a column across a cleaned dataset
138
139    Parameters
140    ----------
141    df : pd.DataFrame
142      The DataFrame containing the desired data.
143    col_name : str
144      The name of the column that the graph will plot.
145    title_name : str
146      The name of the column as you would like to have as the title (e.g. mhm_Genus could be just Genus in the title)
147    """
148    df = df.copy()
149    df[col_name] = df[col_name].replace(-9999, -5)
150
151    counts = df.groupby(col_name).size()
152    title = f"{title_name} Distribution (with Null)"
153    plt.figure(figsize=(10, 5))
154    plt.title(title)
155    plt.ylabel(f"{title_name} Entries  (Log Scale)")
156    plt.yscale("log")
157    plt.bar(counts.index, counts, color="#b30000")
158
159
160def completeness_histogram(df, protocol, completeness_col, completeness_type):
161    """
162    Plots a histogram of the completeness score distribution.
163
164    Parameters
165    ----------
166    df : pd.DataFrame
167      The DataFrame containing the desired data.
168    protocol : str
169      The name of the protocol that the graph will plot.
170    completeness_col : str
171      The column containing the desired completeness metric
172    completness_type : str
173      The type of completeness score (Sub or Cumulative)
174    """
175    plt.figure(figsize=(10, 4))
176    title = f"{protocol} -- {completeness_type} Completeness Scores Frequency Histogram"
177    plt.title(title)
178    plt.hist(df[completeness_col], color="pink", label=completeness_type)
179    plt.xlabel("Scores")
180    plt.ylabel("Count")
181
182
183def save_stored_plots():
184    """
185    Saves any generated graphs currently stored by the pyplot object.
186    """
187    for num in plt.get_fignums():
188        plt.figure(num)
189        title = plt.gca().get_title()
190        plt.savefig(f"{title}.png")