Text and formatting

app.py

@@ -33,6 +33,14 @@ from text_sections import (
     enter_count_data_text,
 )
 
+st.markdown("""
+<style>
+.small-font {
+    font-size:0.85em !important;
+}
+</style>
+""", unsafe_allow_html=True)
+
 
 # Create temporary directory
 
@@ -103,7 +111,9 @@ if input_choice == "Upload predictions":
                 st.session_state["step"] = 1
             else:
                 st.session_state["step"] = 0
-                st.markdown("Please upload a file first (or **generate** some random data to try the function).")
+                st.markdown(
+                    "Please upload a file first (or **generate** some random data to try the function)."
+                )
 
     if st.session_state["step"] >= 1:
         # Read and store (tmp) data
@@ -122,8 +132,8 @@ if input_choice == "Upload predictions":
 elif input_choice == "Upload counts":
     with st.form(key="data_form"):
         upload_counts_text()
-        data_path = st.file_uploader("Upload a dataset", type=["csv"])
-        if st.form_submit_button(label="Use data"):
+        data_path = st.file_uploader("Upload your counts", type=["csv"])
+        if st.form_submit_button(label="Use counts"):
             if data_path:
                 st.session_state["step"] = 1
             else:
@@ -139,9 +149,7 @@ elif input_choice == "Upload counts":
             prediction_col = st.selectbox(
                 "Predictions column", options=list(df.columns)
             )
-            n_col = st.selectbox(
-                "Counts column", options=list(df.columns)
-            )
+            n_col = st.selectbox("Counts column", options=list(df.columns))
 
             if st.form_submit_button(label="Set columns"):
                 st.session_state["step"] = 2
@@ -286,7 +294,7 @@ if st.session_state["step"] >= 2:
         col1, col2, col3 = st.columns([2, 2, 2])
         with col2:
             st.write(df.head(5))
-            st.write(f"{df.shape} (first 5 rows).")
+            st.write(f"{df.shape} (Showing first 5 rows)")
 
     else:
         st.session_state["entered_counts"].to_csv(data_store_path)
@@ -314,7 +322,10 @@ if st.session_state["step"] >= 2:
                 "of another class is excluded.",
             )
         with col2:
-            if st.session_state["input_type"] == "data" and predictions_are_probabilities:
+            if (
+                st.session_state["input_type"] == "data"
+                and predictions_are_probabilities
+            ):
                 prob_of_class = st.selectbox(
                     "Probabilities are of (not working)",
                     options=st.session_state["classes"],
@@ -401,7 +412,7 @@ if st.session_state["step"] >= 2:
         with col3:
             dpi = st.number_input("DPI (not working)", value=320)
 
-        if st.form_submit_button(label="Apply"):
+        if st.form_submit_button(label="Generate plot"):
             st.session_state["step"] = 3
 
     if st.session_state["step"] >= 3:

cvms_version.R

@@ -1 +1 @@
-print(packageVersion("cvms"))
+print(packageVersion("cvms"))

generate_data.R

@@ -3,25 +3,32 @@ library(optparse)
 library(cvms)
 
 option_list <- list(
-    make_option(c("--out_path"), type="character", 
-                help="Path to save data at."),
-    make_option(c("--num_classes"), type="integer", 
-                help="Number of classes."),
-    make_option(c("--num_observations"), type="integer",
-                help="Number of observations."),
-    make_option(c("--seed"), type="integer",
-                help="Number of observations.")
-                
+  make_option(c("--out_path"),
+    type = "character",
+    help = "Path to save data at."
+  ),
+  make_option(c("--num_classes"),
+    type = "integer",
+    help = "Number of classes."
+  ),
+  make_option(c("--num_observations"),
+    type = "integer",
+    help = "Number of observations."
+  ),
+  make_option(c("--seed"),
+    type = "integer",
+    help = "Number of observations."
+  )
 )
- 
-opt_parser <- OptionParser(option_list=option_list)
+
+opt_parser <- OptionParser(option_list = option_list)
 opt <- parse_args(opt_parser)
 
 print(opt)
 
 # Set seed if given
-if (!is.null(opt$seed)){
-    set.seed(opt$seed)
+if (!is.null(opt$seed)) {
+  set.seed(opt$seed)
 }
 
 # Make fairly certain predictions
@@ -31,16 +38,16 @@ rcertain <- function(n) {
 
 # Generate data
 data <- cvms::multiclass_probability_tibble(
-  num_classes=opt$num_classes,
-  num_observations=opt$num_observations,
+  num_classes = opt$num_classes,
+  num_observations = opt$num_observations,
   apply_softmax = TRUE,
   FUN = rcertain,
   class_name = "c",
   add_predicted_classes = TRUE,
   add_targets = TRUE
-) 
+)
 
 data <- data[, c("Predicted Class", "Target")]
 
 # Write to disk
-write.csv(data, file = opt$out_path, row.names=FALSE)
+write.csv(data, file = opt$out_path, row.names = FALSE)

plot.R

@@ -3,59 +3,107 @@ library(optparse)
 library(cvms)
 library(dplyr)
 library(ggplot2)
- 
+
 option_list <- list(
-    make_option(c("--data_path"), type="character", 
-                help="Path to data file (.csv)."),
-    make_option(c("--out_path"), type="character", 
-                help="Path to save confusion matrix plot at."),
-    make_option(c("--data_are_counts"), action="store_true", default=FALSE,
-                help="Indicates that `--data_path` contains counts, not predictions."),
-    make_option(c("--target_col"), type="character", 
-                help="Target column"),
-    make_option(c("--prediction_col"), type="character",
-                help="Prediction column"),
-    make_option(c("--n_col"), type="character",
-                help="Count column (when `--data_are_counts`)."),
-    make_option(c("--classes"), type="character", 
-                help="Comma-separated class names. Only these classes will be used - in the specified order."),
-    make_option(c("--prob_of_class"), type="character", 
-                help="Name of class that probabilities are of."),
-    make_option(c("--palette"), type="character", 
-                help="Color palette."),
-    make_option(c("--width"), type="integer", 
-                help="Width of plot in pixels."),
-    make_option(c("--height"), type="integer", 
-                help="Height of plot in pixels."),
-    make_option(c("--dpi"), type="integer", 
-                help="DPI of plot."),
-    make_option(c("--add_sums"), action="store_true", default=FALSE,
-                help="Wether to add sum tiles."),
-    make_option(c("--add_counts"), action="store_true", default=FALSE,
-                help="Wether to add counts."),
-    make_option(c("--add_normalized"), action="store_true", default=FALSE,
-                help="Wether to add normalized counts (i.e. percentages)."),
-    make_option(c("--add_row_percentages"), action="store_true", default=FALSE,
-                help="Wether to add row percentages."),
-    make_option(c("--add_col_percentages"), action="store_true", default=FALSE,
-                help="Wether to add column percentages."),
-    make_option(c("--add_zero_percentages"), action="store_true", default=FALSE,
-                help="Wether to add percentages to zero-tiles."),
-    make_option(c("--add_zero_text"), action="store_true", default=FALSE,
-                help="Wether to add text to zero-tiles."),
-    make_option(c("--add_zero_shading"), action="store_true", default=FALSE,
-                help="Wether to add shading to zero-tiles."),
-    make_option(c("--add_arrows"), action="store_true", default=FALSE,
-                help="Wether to add arrows to row/sum percentages. Requires additional packages."),
-    make_option(c("--counts_on_top"), action="store_true", default=FALSE,
-                help="Wether to have the counts on top and normalized counts below."),
-    make_option(c("--diag_percentages_only"), action="store_true", default=FALSE,
-                help="Wether to only show diagonal row/column percentages."),
-    make_option(c("--digits"), type="integer",
-                help="Number of digits to show for percentages.")
+    make_option(c("--data_path"),
+        type = "character",
+        help = "Path to data file (.csv)."
+    ),
+    make_option(c("--out_path"),
+        type = "character",
+        help = "Path to save confusion matrix plot at."
+    ),
+    make_option(c("--data_are_counts"),
+        action = "store_true", default = FALSE,
+        help = "Indicates that `--data_path` contains counts, not predictions."
+    ),
+    make_option(c("--target_col"),
+        type = "character",
+        help = "Target column"
+    ),
+    make_option(c("--prediction_col"),
+        type = "character",
+        help = "Prediction column"
+    ),
+    make_option(c("--n_col"),
+        type = "character",
+        help = "Count column (when `--data_are_counts`)."
+    ),
+    make_option(c("--classes"),
+        type = "character",
+        help = "Comma-separated class names. Only these classes will be used - in the specified order."
+    ),
+    make_option(c("--prob_of_class"),
+        type = "character",
+        help = "Name of class that probabilities are of."
+    ),
+    make_option(c("--palette"),
+        type = "character",
+        help = "Color palette."
+    ),
+    make_option(c("--width"),
+        type = "integer",
+        help = "Width of plot in pixels."
+    ),
+    make_option(c("--height"),
+        type = "integer",
+        help = "Height of plot in pixels."
+    ),
+    make_option(c("--dpi"),
+        type = "integer",
+        help = "DPI of plot."
+    ),
+    make_option(c("--add_sums"),
+        action = "store_true", default = FALSE,
+        help = "Wether to add sum tiles."
+    ),
+    make_option(c("--add_counts"),
+        action = "store_true", default = FALSE,
+        help = "Wether to add counts."
+    ),
+    make_option(c("--add_normalized"),
+        action = "store_true", default = FALSE,
+        help = "Wether to add normalized counts (i.e. percentages)."
+    ),
+    make_option(c("--add_row_percentages"),
+        action = "store_true", default = FALSE,
+        help = "Wether to add row percentages."
+    ),
+    make_option(c("--add_col_percentages"),
+        action = "store_true", default = FALSE,
+        help = "Wether to add column percentages."
+    ),
+    make_option(c("--add_zero_percentages"),
+        action = "store_true", default = FALSE,
+        help = "Wether to add percentages to zero-tiles."
+    ),
+    make_option(c("--add_zero_text"),
+        action = "store_true", default = FALSE,
+        help = "Wether to add text to zero-tiles."
+    ),
+    make_option(c("--add_zero_shading"),
+        action = "store_true", default = FALSE,
+        help = "Wether to add shading to zero-tiles."
+    ),
+    make_option(c("--add_arrows"),
+        action = "store_true", default = FALSE,
+        help = "Wether to add arrows to row/sum percentages. Requires additional packages."
+    ),
+    make_option(c("--counts_on_top"),
+        action = "store_true", default = FALSE,
+        help = "Wether to have the counts on top and normalized counts below."
+    ),
+    make_option(c("--diag_percentages_only"),
+        action = "store_true", default = FALSE,
+        help = "Wether to only show diagonal row/column percentages."
+    ),
+    make_option(c("--digits"),
+        type = "integer",
+        help = "Number of digits to show for percentages."
+    )
 )
 
-opt_parser <- OptionParser(option_list=option_list)
+opt_parser <- OptionParser(option_list = option_list)
 opt <- parse_args(opt_parser)
 
 print(opt)
@@ -66,129 +114,144 @@ data_are_counts <- opt$data_are_counts
 target_col <- stringr::str_squish(opt$target_col)
 target_col <- stringr::str_replace_all(target_col, " ", ".")
 prediction_col <- stringr::str_squish(opt$prediction_col)
-prediction_col  <- stringr::str_replace_all(prediction_col, " ", ".")
+prediction_col <- stringr::str_replace_all(prediction_col, " ", ".")
 
 n_col <- NULL
-if (!is.null(opt$n_col)){
+if (!is.null(opt$n_col)) {
     n_col <- stringr::str_squish(opt$n_col)
-    n_col  <- stringr::str_replace_all(n_col, " ", ".")
+    n_col <- stringr::str_replace_all(n_col, " ", ".")
 }
 
 # Read and prepare data frame
-df <- tryCatch({
-    read.csv(opt$data_path)
-}, error=function(e){
-    print(paste0("Failed to read data from ", opt$data_path))
-    print(e)
-    stop(e)
-})
+df <- tryCatch(
+    {
+        read.csv(opt$data_path)
+    },
+    error = function(e) {
+        print(paste0("Failed to read data from ", opt$data_path))
+        print(e)
+        stop(e)
+    }
+)
 print(df)
 
 df <- dplyr::as_tibble(df)
 print(df)
 df[[target_col]] <- as.character(df[[target_col]])
 
-if (isTRUE(data_are_counts)){
+if (isTRUE(data_are_counts)) {
     df[[prediction_col]] <- as.character(df[[prediction_col]])
 }
 
-# Predictions can be either probabilities or 
+# Predictions can be either probabilities or
 # hard class predictions
-if (is.integer(df[[prediction_col]]) || !is.numeric(df[[prediction_col]])){
-    all_present_classes <- sort( 
-        c(unique(df[[target_col]]), 
-          unique(df[[prediction_col]])
+if (is.integer(df[[prediction_col]]) || !is.numeric(df[[prediction_col]])) {
+    all_present_classes <- sort(
+        c(
+            unique(df[[target_col]]),
+            unique(df[[prediction_col]])
         )
     )
 } else {
-    all_present_classes <- sort( 
+    all_present_classes <- sort(
         unique(df[[target_col]])
     )
 }
 
 
-if (!is.null(opt$classes)){
-    classes <- as.character(unlist(strsplit(opt$classes,"[,:]")), recursive=TRUE)
+if (!is.null(opt$classes)) {
+    classes <- as.character(
+        unlist(strsplit(opt$classes, "[,:]")),
+        recursive = TRUE
+    )
 } else {
     classes <- all_present_classes
 }
-print(paste0("Selected Classes: ", paste0(classes, collapse=", ")))
+print(paste0("Selected Classes: ", paste0(classes, collapse = ", ")))
 
-if (!isTRUE(data_are_counts)){
-    # We remove the unwanted classes from the confusion matrix 
+if (!isTRUE(data_are_counts)) {
+    # We remove the unwanted classes from the confusion matrix
     # (easier - possibly slower in edge cases)
     family <- ifelse(length(all_present_classes) == 2, "binomial", "multinomial")
     print(df)
 
     # TODO : use prob_of_class to ensure probabilities are interpreted correctly!!
     # Might need to invert them to get it to work!
-    evaluation <- tryCatch({
-        cvms::evaluate(
-            data=df,
-            target_col=target_col,
-            prediction_cols=prediction_col,
-            type=family,
-        )
-    }, error=function(e){
-        print("Failed to evaluate data.")
-        print(head(df, 5))
-        print(e)
-        stop(e)
-    })
+    evaluation <- tryCatch(
+        {
+            cvms::evaluate(
+                data = df,
+                target_col = target_col,
+                prediction_cols = prediction_col,
+                type = family,
+            )
+        },
+        error = function(e) {
+            print("Failed to evaluate data.")
+            print(head(df, 5))
+            print(e)
+            stop(e)
+        }
+    )
 
     confusion_matrix <- evaluation[["Confusion Matrix"]][[1]]
-
 } else {
     confusion_matrix <- dplyr::rename(
-        df, 
+        df,
         Target = !!target_col,
         Prediction = !!prediction_col,
         N = !!n_col
     )
 }
 
-confusion_matrix  <- dplyr::filter(
+confusion_matrix <- dplyr::filter(
     confusion_matrix,
     Prediction %in% classes,
     Target %in% classes
 )
 
 
-confusion_matrix_plot <- tryCatch({
-    cvms::plot_confusion_matrix(
-        confusion_matrix, 
-        class_order=classes,
-        add_sums=opt$add_sums,
-        add_counts=opt$add_counts,
-        add_normalized=opt$add_normalized,
-        add_row_percentages=opt$add_row_percentages,
-        add_col_percentages=opt$add_col_percentages,
-        rm_zero_percentages=!opt$add_zero_percentages,
-        rm_zero_text=!opt$add_zero_text,
-        add_zero_shading=opt$add_zero_shading,
-        add_arrows=opt$add_arrows,
-        counts_on_top=opt$counts_on_top,
-        diag_percentages_only=opt$diag_percentages_only,
-        digits=as.integer(opt$digits),
-        palette=opt$palette
-    )
-}, error=function(e){
-    print("Failed to create plot from confusion matrix.")
-    print(confusion_matrix)
-    print(e)
-    stop(e)
-})
-
-tryCatch({
-    ggplot2::ggsave(
-        opt$out_path, 
-        width=opt$width, 
-        height=opt$height, 
-        dpi=opt$dpi, 
-        units="px"
-    )
-}, error=function(e){
-    print(paste0("Failed to ggsave plot to: ", opt$out_path))
-    print(e)
-    stop(e)
-})
+confusion_matrix_plot <- tryCatch(
+    {
+        cvms::plot_confusion_matrix(
+            confusion_matrix,
+            class_order = classes,
+            add_sums = opt$add_sums,
+            add_counts = opt$add_counts,
+            add_normalized = opt$add_normalized,
+            add_row_percentages = opt$add_row_percentages,
+            add_col_percentages = opt$add_col_percentages,
+            rm_zero_percentages = !opt$add_zero_percentages,
+            rm_zero_text = !opt$add_zero_text,
+            add_zero_shading = opt$add_zero_shading,
+            add_arrows = opt$add_arrows,
+            counts_on_top = opt$counts_on_top,
+            diag_percentages_only = opt$diag_percentages_only,
+            digits = as.integer(opt$digits),
+            palette = opt$palette
+        )
+    },
+    error = function(e) {
+        print("Failed to create plot from confusion matrix.")
+        print(confusion_matrix)
+        print(e)
+        stop(e)
+    }
+)
+
+tryCatch(
+    {
+        ggplot2::ggsave(
+            opt$out_path,
+            width = opt$width,
+            height = opt$height,
+            dpi = opt$dpi,
+            units = "px"
+        )
+    },
+    error = function(e) {
+        print(paste0("Failed to ggsave plot to: ", opt$out_path))
+        print(e)
+        stop(e)
+    }
+)

text_sections.py

@@ -23,34 +23,59 @@ def intro_text():
     col1, col2 = st.columns([8, 2])
     with col1:
         st.title("Plot Confusion Matrix")
-        st.write(
-            "This application allows you to plot a confusion matrix based on your own data. "
+        st.markdown(
+            "A confusion matrix plot is a great tool for inspecting your "
+            "machine learning model's performance on a classification task. "
+            "This application enables you to plot a confusion matrix on your own data, "
+            "**without a single line of code**. \n\n"
+            "It's designed for high flexibility AND quick results with good default settings.\n\n"
         )
     with col2:
         st.image(
             "https://github.com/LudvigOlsen/cvms/raw/master/man/figures/cvms_logo_242x280_250dpi.png",
             width=125,
         )
-
+    st.markdown("""---""")
+    col1, col2 = st.columns(2)
+    with col1:
+        st.subheader("Have your data ready?")
+        st.markdown(  # TODO: Make A,B, etc. icons
+            "Upload a csv file with either: \n\n"
+            "A) **Targets** and **predictions**. \n\n"
+            "B) Existing confusion matrix **counts**. \n\n"
+            "--> Specify the columns to use.\n\n"
+            "--> Press **Generate plot**.\n\n"
+        )
+    with col2:
+        st.subheader("No data to upload?")
+        st.markdown(
+            "No worries! Either: \n\n"
+            "C) **Input** your counts directly! \n\n"
+            "D) **Generate* some data with **very** easy controls! \n\n"
+            "--> Press **Generate plot**.\n\n"
+        )
+    st.markdown("""---""")
     st.write(
         "The plot is created with the [**cvms**](https://github.com/LudvigOlsen/cvms) R package "
         f"(v/{get_cvms_version()}, LR Olsen & HB Zachariae, 2019)."
-    )
+    )  # TODO Add citation stuff
 
-    st.write(
-        "DATA PRIVACY: In order to transfer the data "
-        "between python and R, it is temporarily stored on the servers. "
-        "While we, the authors, have no intention of looking at your data, we make "
-        "*no guarantees* about the privacy of your data (it is not our servers). "
+    st.markdown(
+        '<p class="small-font">'
+        "DATA PRIVACY: For technical reasons, the uploaded data is temporarily stored "
+        "on the server. While we, the authors, won't access your data, we make "
+        "*no guarantees* about the privacy of your data (not our servers). "
         "Please do not upload sensitive data. The application "
-        "only requires columns with predictions and targets."
+        "only requires either predictions and targets or counts. "
+        "</p>",
+        unsafe_allow_html=True,
     )
 
 
 def generate_data_text():
     st.subheader("Generate data")
     st.write(
-        "If you just want to try out the application, you can generate a dataset with targets and predictions. "
+        "Quickly try the application by generating a dataset with targets and predictions. "
         "Select a number of classes and observations, and you're ready to go! "
     )
 
@@ -77,6 +102,7 @@ def upload_counts_text():
         "See example of such a .csv file [here] (TODO). "
     )
 
+
 def upload_predictions_text():
     st.subheader("Upload your predictions")
     st.markdown(
@@ -85,7 +111,7 @@ def upload_predictions_text():
         "Targets will be converted into strings. \n\n"
         "2) A `prediction` column.  \n"
         "Predictions can be probabilities (binary classification only) or class predictions. \n\n"
-        "Other columns are currently ignored.  \n\n" 
+        "Other columns are currently ignored.  \n\n"
         "You will have the option to select the names of these two columns, so don't "
         "worry too much about the column names in the uploaded data."
     )