I'm trying to create a line chart which is based on filter option along with data table. If I click the filter option it only changed in data table. But I want to set the filters for both Line chart and data table. Kindly help me.
My server.r code is:
output$grp_stacked_bar <- renderPlotly({
data <- ex
agg_sum <- aggregate(data$Result,by=list(Category = data$Donor_ID),FUN=sum, na.rm=TRUE)
p <- plot_ly(
data = agg_sum,
x = ~Category,
y = ~x,
type = "scatter",
mode = "lines+markers"
) %>% layout(title = "Functional Outlier Details", xaxis = list(title = "Donor_ID"),
yaxis = list(title = "Result"))
p
})
ui.r:
column(width = 10,
fluidRow(h2("Function Outlier Details",class="box-title",align="center"),
plotlyOutput(height="48vh",width ="82vw",outputId = "grp_stacked_bar")%>%withSpinner(color="#0dc5c1",hide.ui = FALSE,image.height = "73px",image.width = "145px",image= ".\\logo_gif2.gif"))
)
My sample dataset is:
enter image description here
Related
I am new to TensorFlow and I get my code running successfully by modifying tutorials from the official website.
I checked some other answers on StackOverflow, which says my problem is likely due to something is being added to the graph every time. However, I have no idea where to look for the code that might have caused this.
Also, I used tf.py_function to map the dataset because I really need to enable eagerly mode in the mapping.
def get_dataset(data_index):
# data_index is a Pandas Dataframe that contains image/label pair info, each row is one pair
data_index = prepare_data_index(data_index)
# shuffle dataframe here because dataset.shuffle is taking very long time.
data_index = data_index.sample(data_index.shape[0])
path = path_to_img_dir
# list of dataframe indices indicating rows that are going to be included in the dataset for training.
indices_ls = ['{}_L'.format(x) for x in list(data_index.index)] + ['{}_R'.format(x) for x in list(data_index.index)]
# around 310k images
image_count = len(indices_ls)
list_ds = tf.data.Dataset.from_tensor_slices(indices_ls)
# dataset.shuffle is commented out because it takes too much time
# list_ds = list_ds.shuffle(image_count, reshuffle_each_iteration=False)
val_size = int(image_count * 0.2)
train_ds = list_ds.skip(val_size)
val_ds = list_ds.take(val_size)
def get_label(index):
index = str(np.array(index).astype(str))
delim = index.split('_')
state = delim[1]
index = int(delim[0])
if state == 'R':
label = data_index.loc[index][right_labels].to_numpy().flatten()
elif state == 'L':
label = data_index.loc[index][left_labels].to_numpy().flatten()
return tf.convert_to_tensor(label , dtype=tf.float16)
def get_img(index):
index = str(np.array(index).astype(str))
delim = index.split('_')
state = delim[1]
index = int(delim[0])
file_path = '{}_{}.jpg'.format(data_index.loc[index, 'sub_folder'],
str(int(data_index.loc[index, 'img_index'])).zfill(4)
)
img = tf.io.read_file(os.path.join(path, file_path))
img = tf.image.decode_jpeg(img, channels=3)
full_width = 320
img = tf.image.resize(img, [height, full_width])
# Crop half of the image depending on the state
if state == 'R':
img = tf.image.crop_to_bounding_box(img, offset_height=0, offset_width=0, target_height=height,
target_width=int(full_width / 2))
img = tf.image.flip_left_right(img)
elif state == 'L':
img = tf.image.crop_to_bounding_box(img, offset_height=0, offset_width=int(full_width / 2), target_height=height,
target_width=int(full_width / 2))
img = tf.image.resize(img, [height, width])
img = tf.keras.preprocessing.image.array_to_img(
img.numpy(), data_format=None, scale=True, dtype=None
)
# Apply auto white balancing, output an np array
img = AWB(img)
img = tf.convert_to_tensor(img, dtype=tf.float16)
return img
def process_path(index):
label = get_label(index)
img = get_img(index)
return img, label
AUTOTUNE = tf.data.experimental.AUTOTUNE
train_ds = train_ds.map(lambda x: tf.py_function(
process_path,
[x], (tf.float16, tf.float16)), num_parallel_calls=AUTOTUNE)
val_ds = val_ds.map(lambda x: tf.py_function(
process_path,
[x], (tf.float16, tf.float16)), num_parallel_calls=AUTOTUNE)
def configure_for_performance(ds):
ds = ds.cache()
# ds = ds.shuffle(buffer_size=image_count)
ds = ds.batch(batch_size)
ds = ds.prefetch(buffer_size=AUTOTUNE)
return ds
train_ds = configure_for_performance(train_ds)
val_ds = configure_for_performance(val_ds)
return train_ds, val_ds
Can anyone please help me? Thanks!
Here is the rest of my code.
def initialize_model():
IMG_SIZE = (height, width)
preprocess_input = tf.keras.applications.vgg19.preprocess_input
IMG_SHAPE = IMG_SIZE + (3,)
base_model = tf.keras.applications.VGG19(input_shape=IMG_SHAPE,
include_top=False,
weights='imagenet')
global_average_layer = tf.keras.layers.GlobalAveragePooling2D()
prediction_layer = tf.keras.layers.Dense(class_num, activation=tf.nn.sigmoid, use_bias=True)
inputs = tf.keras.Input(shape=(height, width, 3))
x = preprocess_input(inputs)
x = base_model(x, training=True)
x = global_average_layer(x)
outputs = prediction_layer(x)
model = tf.keras.Model(inputs, outputs)
def custom_loss(y_gt, y_pred):
L1_loss_out = tf.math.abs(tf.math.subtract(y_gt, y_pred))
scaler = tf.pow(50.0, y_gt)
scaled_loss = tf.math.multiply(L1_loss_out, scaler)
scaled_loss = tf.math.reduce_mean(
scaled_loss, axis=None, keepdims=False, name=None
)
return scaled_loss
base_learning_rate = 0.001
model.compile(optimizer=tf.keras.optimizers.SGD(learning_rate=base_learning_rate, momentum=0.9),
loss=custom_loss,
metrics=['mean_absolute_error']
)
return model
def train(data_index, epoch_num, save_path):
train_dataset, validation_dataset = get_dataset(data_index)
model = initialize_model()
model.summary()
history = model.fit(train_dataset,
epochs=epoch_num,
validation_data=validation_dataset)
model.save_weights(save_path)
return model, history
I am using amcharts line chart.
I have data for last 24 hours and its recorded for every seconds.
I am trying to group data in amcharts but it displays only 2 data points on chart.
1 data point is from yesterday and 1 from today.
Here is my code:
var multiLineChart = am4core.create(
"multilineChartdiv",
am4charts.XYChart
);
multiLineChart.paddingRight = 20;
multiLineChart.data = historicalData;
var dateAxis1 = multiLineChart.xAxes.push(new am4charts.DateAxis());
dateAxis1.renderer.grid.template.location = 0;
dateAxis1.minZoomCount = 1;
dateAxis1.renderer.minGridDistance = 60;
// dateAxis1.baseInterval = {
// timeUnit: "minute",
// count: 5,
// };
// this makes the data to be grouped
dateAxis1.groupData = true;
dateAxis1.groupCount = 500;
var valueAxis = multiLineChart.yAxes.push(new am4charts.ValueAxis());
var series1 = multiLineChart.series.push(new am4charts.LineSeries());
series1.dataFields.dateX = "date";
series1.dataFields.valueY = "heartRate";
series1.tooltipText = "{valueY}";
series1.tooltip.pointerOrientation = "vertical";
series1.tooltip.background.fillOpacity = 0.5;
multiLineChart.cursor = new am4charts.XYCursor();
multiLineChart.cursor.xAxis = dateAxis1;
var scrollbarX = new am4core.Scrollbar();
scrollbarX.marginBottom = 20;
multiLineChart.scrollbarX = scrollbarX;
I need to show data points for at least every 5 or 10 minutes.
If your timestamp is a string, make sure the inputDateFormat is set to match your date format as documented here as the default format is yyyy-MM-dd, truncating everything else to look like daily data, similar to your screenshot:
chart.dateFormatter.inputDateFormat = 'yyyy-MM-dd HH:mm:ss' //adjust as needed
Since your data is in seconds, it is also recommended to set the baseInterval accordingly to also ensure that your data is rendered correctly.
dateAxis1.baseInterval = {
timeUnit: "second",
count: 1,
};
From what I can tell in my limited review of bokeh documentation the ability to click on a glyph on a plot then present a Dialog box or Datatable is a feature not yet available. I do not want the Datatable to be presented until a glyph has been selected. Would ideally like the ability to hide the Dialog or Datatable as well.
It seems that bokeh.models.widgets.dialog were deprecated sometime after 0.10.0. I could use that but its not available in python 3.7 at this point. Suggestions?
Some features are not officially supported but sometimes one can come up with a work-arounds like this one (tested on Bokeh v1.0.4):
from bokeh.plotting import figure, show
from bokeh.layouts import column
from bokeh.models import ColumnDataSource, Slider, DataTable, TableColumn, CustomJS
plot = figure(tools = 'tap')
source = ColumnDataSource(dict(x = list(range(6)), y = [x ** 2 for x in range(6)]))
circles = plot.circle('x', 'y', source = source, size = 20)
slider = Slider(start = -1, end = 5, value = 6, step = 1, title = "i", width = 300)
columns = [TableColumn(field = "x", title = "x"), TableColumn(field = "y", title = "x**2")]
table = DataTable(source = source, columns = columns, width = 320)
plot.js_on_event('tap', CustomJS(args = {'table': table, 'source': source, 'slider': slider}, code = '''
const selected_index = source.selected.indices[0]
if (selected_index != null)
table.height = 0;
else
table.height = slider.value * 25 + 25;'''))
callback_code = """ i = slider.value;
new_data = {"x": [0,1,2,3,4,5], "y": [0,1,4,9,16,25]}
table.source.data = new_data
table.height = i * 25 + 25; """
callback = CustomJS(args = dict(slider = slider, table = table), code = callback_code)
slider.js_on_change('value', callback)
show(column(slider, plot, table))
Result:
When using officer and the rvg graphics I want my images to not have the background white square.
The reason for this is that some functions leave a lot of empty space up at the top and often this will overlap with the header. I can go in and delete each one but when making 20 or more plots this gets tedious fast.
Specifically I want to use this with the forest function in metafor used here in the example.
Before deleting the background.
**
After deleting the background.
**
library(officer)
doc <- read_pptx()
doc <- add_slide(doc, "Title and Content", "Office Theme")
doc <- ph_with_vg_at(doc, code = barplot(1:5, col = 2:6),
left = 1, top = 2, width = 6, height = 4)
library(metafor)
library(gemtc)
network <-mtc.network(smoking$data.ab)
data(dat.bcg)
plot(network)
res <- rma(ai=tpos, bi=tneg, ci=cpos, di=cneg, data=dat.bcg, measure="RR",
slab=paste(author, year, sep=", "), method="REML")
forest(res)
doc <- add_slide(doc, "Title and Content", "Office Theme")
doc <- ph_with_vg(doc, code = plot(network), type = "body", width = 8, height = 6)
doc <- add_slide(doc, "Title and Content", "Office Theme")
doc <- ph_with_vg(doc, code = forest(res), type = "body", width = 8, height = 6)
print(doc, target = "vg.pptx")
This is now possible using the latest rvg update and setting the background to "NA".
anyplot = dml(code = barplot(1:5, col = 2:6), bg = "NA")
doc <- read_pptx()
doc <- add_slide(doc)
doc <- ph_with(doc, anyplot, location = ph_location_fullsize())
print(doc, target = "bg.pptx")
I am trying to build a application in R shiny, where I am using a handsontable to take inputs from users. One column in my handsontable is having dropdowns where I need multiple selection from the user.
For example in my below sample code, I want to allow user to select multiple values in 'big' column (i.e. user should be able to select A,B,C for first row and likewise for other rows)
library(shiny)
library(rhandsontable)
shinyApp(
ui = fluidPage(
fluidRow(
column(12,
sliderInput('input', label = "Rows",
min = 1, max = nrow(iris), value = 10)
),
column(12
,
rHandsontableOutput('table')
)
)
),
server = function(input, output) {
DF = data.frame(val = 1:10, bool = TRUE, big = LETTERS[1:10],
small = letters[1:10],
dt = seq(from = Sys.Date(), by = "days", length.out = 10),
stringsAsFactors = FALSE)
# try updating big to a value not in the dropdown
output$table <- renderRHandsontable(
rhandsontable(DF, rowHeaders = NULL, width = 550, height = 300) %>%
hot_col(col = "big", type = "dropdown", source = LETTERS) %>%
hot_col(col = "small", type = "autocomplete", source = letters,
strict = FALSE)
)
}
)
Let me know if anyone has faced the same problem and resolved the same.
Your question is a bit confusing for me, you want that the slider reacts to how many selected options do the user put or...?
Then I don't know why you put max = nrow(iris) if you are not using the iris dataframe, that makes no sense.
Finally if you want to have a reactive UI so when the user put something, the ui change, you will need this:
1- The reactive function for server:
`outSli <- reactive({
##here you put your action
})`
2- The observe function for server to change the ui slider:
`observe({
##here you put the outSli() that contains your action or algorithm and then you update the slider
updateSliderInput(session, "input", label = NULL, value = NULL,
min = NULL, max = NULL, step = NULL)
##all those NULLs is what you change with your outSli()
})`
I hope it helps.