import os
import uuid
from typing import Callable
import pandas as pd
import streamlit as st
from .filter import filter_data_using_slider
from .metricplot import render_metric_plot
from .resulttable import configure_aggrid, prepare_display_dataframe, render_aggrid
[docs]
def initialize_main_slider(slider_id_uuid: str, default_val_slider: float) -> None:
"""
Initialize the slider for the main data.
We use a slider uuid and associate a defalut value with it.
- self.variables_quant.slider_id_uuid
- self.variables_quant.default_val_slider
"""
key = slider_id_uuid
if key not in st.session_state.keys():
st.session_state[key] = uuid.uuid4()
_id_of_key = st.session_state[key]
if _id_of_key not in st.session_state.keys():
st.session_state[_id_of_key] = default_val_slider
[docs]
def generate_main_slider(slider_id_uuid: str, description_slider_md: str, default_val_slider: float) -> None:
"""
Create a slider input.
"""
# key for slider_uuid in session state
if slider_id_uuid not in st.session_state:
st.session_state[slider_id_uuid] = uuid.uuid4()
slider_key = st.session_state[slider_id_uuid]
fpath = description_slider_md
st.markdown(open(fpath, "r").read())
default_value = st.session_state.get(slider_key, default_val_slider)
st.select_slider(
label="Minimal precursor quantifications (# samples)",
options=[1, 2, 3, 4, 5, 6],
value=default_value,
key=slider_key,
)
[docs]
def generate_main_selectbox(variables_quant, selectbox_id_uuid) -> None:
"""
Create the selectbox for the Streamlit UI.
"""
if selectbox_id_uuid not in st.session_state.keys():
st.session_state[selectbox_id_uuid] = uuid.uuid4()
try:
# TODO: Other labels based on different modules, e.g. mass tolerances are less relevant for DIA
st.selectbox(
"Select label to plot",
variables_quant.metric_plot_labels,
key=st.session_state[selectbox_id_uuid],
)
except Exception as e:
st.error(f"Unable to create the selectbox: {e}", icon="🚨")
[docs]
def display_download_section(variables_quant, reset_uuid=False) -> None:
"""
Render the selector and area for raw data download.
Parameters
----------
reset_uuid : bool, optional
Whether to reset the UUID, by default False.
"""
if len(st.session_state[variables_quant.all_datapoints]) == 0:
st.error("No data available for download.", icon="🚨")
return
downloads_df = st.session_state[variables_quant.all_datapoints][["id", "intermediate_hash"]]
downloads_df.set_index("intermediate_hash", drop=False, inplace=True)
if variables_quant.placeholder_downloads_container not in st.session_state.keys() or reset_uuid:
st.session_state[variables_quant.placeholder_downloads_container] = st.empty()
st.session_state[variables_quant.download_selector_id_uuid] = uuid.uuid4()
# with st.session_state[variables_quant.placeholder_downloads_container].container(border=True):
st.subheader("Download raw datasets")
# Sort the intermediate_hash values and get the corresponding ids
sorted_indices = sorted(range(len(downloads_df["id"])), key=lambda i: downloads_df["id"].iloc[i])
sorted_intermediate_hash = [downloads_df["intermediate_hash"].iloc[i] for i in sorted_indices]
sorted_ids = [downloads_df["id"].iloc[i] for i in sorted_indices]
st.selectbox(
"Select dataset",
sorted_intermediate_hash,
index=None,
key=st.session_state[variables_quant.download_selector_id_uuid],
format_func=lambda x: sorted_ids[sorted_intermediate_hash.index(x)],
)
if (
st.session_state[st.session_state[variables_quant.download_selector_id_uuid]] is not None
and st.secrets["storage"]["dir"] is not None
):
dataset_path = (
st.secrets["storage"]["dir"]
+ "/"
+ st.session_state[st.session_state[variables_quant.download_selector_id_uuid]]
)
if os.path.isdir(dataset_path):
files = os.listdir(dataset_path)
for file_name in files:
path_to_file = dataset_path + "/" + file_name
with open(path_to_file, "rb") as file:
st.download_button(file_name, file, file_name=file_name)
else:
st.write(
"Directory for this dataset does not exist, this should not happen"
" on the server, but is expected in the local development environment."
)
[docs]
def display_existing_results(variables_quant, ionmodule) -> None:
"""
Orchestrates the full display of quantification results in Streamlit,
including plotting and interactive tabular output with styling.
Parameters
----------
variables_quant : object
Object containing quantification variables including data points,
slider/selectbox UUIDs, and configuration flags.
ionmodule : object
Module responsible for filtering and transforming ion data.
"""
initialize_and_filter_data(variables_quant, ionmodule)
data_points_filtered = variables_quant.filtered_data
metric = display_metric_selector(variables_quant)
# ROC-AUC has no mode variants (it's already species-aware by design)
if metric == "ROC-AUC":
mode = None
else:
mode = display_metric_calc_approach_selector(variables_quant)
# prepare plot key explicitly for tab 1
key = variables_quant.result_plot_uuid
if key not in st.session_state.keys():
st.session_state[key] = uuid.uuid4()
_id_of_key = st.session_state[key]
highlight_point_id = render_metric_plot(
data_points_filtered,
metric,
mode,
label=st.session_state[st.session_state[variables_quant.selectbox_id_uuid]],
key=_id_of_key,
)
df_display = prepare_display_dataframe(data_points_filtered, highlight_point_id)
grid_options = configure_aggrid(df_display)
# prepare df key explicitly for tab 1
key = variables_quant.table_id_uuid
if key not in st.session_state.keys():
st.session_state[key] = uuid.uuid4()
_id_of_key = st.session_state[key]
render_aggrid(df_display, grid_options, key=_id_of_key)
offer_download(df_display)
display_download_section(variables_quant=variables_quant)
# === Modular Functions ===
[docs]
def initialize_and_filter_data(variables_quant, ionmodule):
initialize_main_data_points(
all_datapoints=variables_quant.all_datapoints,
obtain_all_data_points=ionmodule.obtain_all_data_points,
)
variables_quant.filtered_data = filter_data_using_slider(
slider_id_uuid=variables_quant.slider_id_uuid,
all_datapoints=variables_quant.all_datapoints,
filter_data_point=ionmodule.filter_data_point,
)
[docs]
def display_metric_selector(variables_quant) -> str:
key = variables_quant.metric_selector_uuid
if key not in st.session_state.keys():
st.session_state[key] = uuid.uuid4()
_id_of_key = st.session_state[key]
# TODO: Add "ROC-AUC" to options list to enable ROC-AUC metric display
return st.radio(
"Select metric to plot",
options=["Median", "Mean"],
help="Toggle between median and mean absolute difference metrics.",
key=_id_of_key,
)
[docs]
def display_metric_calc_approach_selector(variables_quant) -> str:
key = variables_quant.metric_calc_approach_selector_uuid
if key not in st.session_state.keys():
st.session_state[key] = uuid.uuid4()
_id_of_key = st.session_state[key]
return st.radio(
"Select metric calculation approach",
options=["Global", "Species-weighted"],
help="Toggle between species-weighted and global absolute difference metrics. Global considers all metrics equally, while species-weighted account for species abundance variations, i.e. the mean/median metric is calculated per species first before averaging across species.",
key=_id_of_key,
)
[docs]
def offer_download(df: pd.DataFrame, filename: str = "quantification_results.csv") -> None:
"""
Adds a download button to export the displayed DataFrame as a CSV file.
Parameters
----------
df : pd.DataFrame
The DataFrame to be downloaded.
filename : str, optional
The name of the file to download, by default "quantification_results.csv".
"""
csv_data = df.to_csv(index=False).encode("utf-8")
st.download_button(label="📥 Download table as CSV", data=csv_data, file_name=filename, mime="text/csv")
[docs]
def initialize_main_data_points(all_datapoints: str, obtain_all_data_points: Callable) -> None:
"""
Initialize the all_datapoints variable in the session state.
"""
if all_datapoints not in st.session_state.keys():
st.session_state[all_datapoints] = None
st.session_state[all_datapoints] = obtain_all_data_points(all_datapoints=st.session_state[all_datapoints])
