Guide

Configuration

Environment variables, config files, feature flags

Overview

This page explains how to configure the plant-disease-predictor application so it runs correctly in your environment. Configuration is split across two files — config.toml and credentials.toml — which control server behavior, logging, browser connectivity, and account credentials. Getting these settings right before you launch ensures the app is reachable, performs well, and behaves predictably when you upload leaf images for disease analysis.

Prerequisites

Before adjusting any configuration, make sure you have the following in place:

Python 3.8 or later installed on your system
pip (Python package manager) available in your terminal
The following Python packages at the exact versions used by this project:
- numpy==1.26.3
- streamlit==1.30.0
- tensorflow==2.15.0.post1
Write access to the project directory so you can edit config.toml and credentials.toml
A terminal or command prompt open at the project root

Installation

Follow these steps to install all dependencies and place the configuration files correctly.

Clone or download the project to your local machine and navigate into the project directory:
```
cd plant-disease-predictor
```
Install the required Python packages using the provided requirements.txt:
```
pip install -r requirements.txt
```
This installs numpy, streamlit, and tensorflow at the exact versions the application was tested with. Installing different versions may cause unexpected behavior.
Verify the configuration files exist in the project root. You should see both:
```
config.toml
credentials.toml
```
If either file is missing, create it manually using the examples in the Configuration section below.
Confirm Streamlit is available by running:
```
streamlit --version
```
Expected output:
```
Streamlit, version 1.30.0
```

Once installation is complete, you are ready to adjust configuration values before launching.

Configuration

All configuration lives in two TOML files at the project root. Edit these files with any text editor before starting the application.

`config.toml`

This file controls Streamlit's runtime, server, logging, and browser behavior.

`[global]`

Key	Default	Valid values	Effect
`showWarningOnDirectExecution`	`true`	`true`, `false`	When `true`, Streamlit prints a warning if you run the script with `python` instead of `streamlit run`. Keep this enabled during development to avoid confusion.

`[logger]`

Key	Default	Valid values	Effect
`level`	`"info"`	`"error"`, `"warning"`, `"info"`, `"debug"`	Controls the verbosity of log output. The current setting is `"debug"`, which prints detailed messages useful for diagnosing prediction or connectivity issues. Switch to `"info"` or `"warning"` in production to reduce noise.

`[runner]`

Key	Default	Valid values	Effect
`magicEnabled`	`true`	`true`, `false`	Allows bare Python expressions on a single line to be rendered in the app. Disabling this is rarely necessary unless you encounter unexpected output in the UI.

`[server]`

These settings control how the application server listens for and handles connections.

Key	Default	Valid values	Effect
`folderWatchBlacklist`	`[]`	List of paths	Folders the server will not watch for file changes. Useful for excluding large virtual-environment directories from hot-reload scanning.
`headless`	`false` (platform-dependent)	`true`, `false`	Set to `true` to prevent the server from opening a browser window automatically on startup. Required when running on a remote server or inside a container.
`liveSave`	`false`	`true`, `false`	When `true`, the app is immediately shared for live monitoring. Leave as `false` for standard local or self-hosted use.
`runOnSave`	`false`	`true`, `false`	When `true`, the app automatically reruns whenever the script file changes on disk. Convenient during development; keep `false` in production.
`port`	`8501`	Any valid port integer	The port the server listens on. The current value is `80` (standard HTTP), which means you can access the app without specifying a port in your browser URL. Ensure this port is open in your firewall or security group.
`enableCORS`	`true`	`true`, `false`	Enables Cross-Origin Resource Sharing protection. Currently set to `false` — re-enable this (`true`) in any internet-facing deployment to prevent unauthorized cross-origin requests.

`[browser]`

These settings tell the browser-side client where to connect.

Key	Default	Valid values	Effect
`serverAddress`	`"localhost"`	IP address or DNS name	The address the browser uses to reach the server. `"0.0.0.0"` means the server accepts connections on all network interfaces, making it reachable from other machines on the network.
`gatherUsageStats`	`true`	`true`, `false`	When `true`, anonymous usage data is sent to Streamlit. Set to `false` if your deployment environment does not permit outbound telemetry.
`serverPort`	Same as `server.port`	Any valid port integer	The port the browser connects to. This must match `server.port` (both are set to `80` here).

`credentials.toml`

This file stores optional account credentials used by Streamlit's sharing features.

`[general]`

Key	Default	Effect
`email`	`""`	Your Streamlit account email. Required only if you use Streamlit Cloud sharing or `liveSave`. Leave blank for self-hosted deployments.

Usage

Once your configuration files are in place and dependencies are installed, you launch the application with a single command.

Starting the application:

streamlit run app.py --config config.toml

Because server.headless is set to true, no browser window opens automatically. Open your browser manually and navigate to:

http://localhost:80

or, if you are accessing from another machine on the network:

http://<your-server-ip>:80

During development, you may want to enable runOnSave so the app reloads whenever you modify the script:

[server]
runOnSave = true

In production, keep runOnSave = false and switch the log level to reduce output:

[logger]
level = "warning"

Restricting external access: If you only need local access, change serverAddress to "localhost" in [browser] and ensure server.port is set to a port not exposed publicly.

Tip: Always restart the application after editing config.toml for changes to take effect.

Examples

The examples below show common configuration scenarios you might encounter when deploying the plant-disease-predictor.

Example 1 — Local development setup (verbose logging, auto-reload)

Use this configuration when you are developing locally and want maximum feedback:

[logger]
level = "debug"

[server]
headless = false
runOnSave = true
port = 8501
enableCORS = true

[browser]
serverAddress = "localhost"
serverPort = 8501

Launch with:

streamlit run app.py

Expected behavior: Streamlit opens a browser tab at http://localhost:8501 automatically. Every time you save app.py, the app reruns. Debug messages appear in your terminal.

Example 2 — Production server deployment (headless, restricted logging)

Use this configuration when hosting on a remote server or VM:

[logger]
level = "warning"

[server]
headless = true
runOnSave = false
port = 80
enableCORS = true

[browser]
serverAddress = "0.0.0.0"
serverPort = 80

Launch with:

streamlit run app.py

Expected behavior: No browser opens on the server. The app is reachable at http://<your-server-ip> from any machine. Only warnings and errors appear in logs.

Example 3 — Disabling usage statistics

If your environment blocks outbound telemetry, add this to config.toml:

[browser]
gatherUsageStats = false

Expected behavior: No usage data is sent to Streamlit's servers. All other functionality remains unchanged.

Troubleshooting

Use the table below to diagnose and fix the most common configuration-related problems.

Problem: The app fails to start because port 80 is already in use

Symptom: Terminal shows OSError: [Errno 98] Address already in use or similar.
Likely cause: Another process (e.g., a web server like Apache or Nginx) is already listening on port 80.
Fix: Either stop the conflicting process, or change server.port and browser.serverPort in config.toml to an available port such as 8501, then restart.

Problem: Navigating to localhost:80 shows a blank page or connection refused

Symptom: Browser cannot load the app after running streamlit run app.py.
Likely cause: server.headless is false and the app tried to open a browser before fully initializing, or browser.serverAddress is set to "0.0.0.0" but you are connecting via localhost.
Fix: Confirm the terminal shows You can now view your Streamlit app in your browser. If serverAddress is "0.0.0.0", try navigating to http://127.0.0.1:80 instead.

Problem: Changes to config.toml have no effect

Symptom: After editing a setting, the app still behaves as before.
Likely cause: The application was not restarted after saving the file.
Fix: Stop the running process (Ctrl+C) and relaunch with streamlit run app.py. Configuration is only read at startup.

Problem: Excessive log output is cluttering the terminal

Symptom: The terminal is flooded with DEBUG messages during normal operation.
Likely cause: logger.level is set to "debug" in config.toml.
Fix: Change the level to "info" or "warning" for less verbose output, then restart the app.

Problem: The app is accessible from unintended external machines

Symptom: Users outside your network can reach the app, or security scans flag an open port.
Likely cause: browser.serverAddress is "0.0.0.0" and enableCORS is false, leaving the server open without cross-origin protection.
Fix: Set enableCORS = true in [server] immediately. If you only need local access, also change browser.serverAddress to "localhost" and restrict port 80 at your firewall level.

Problem: tensorflow or numpy import errors on startup

Symptom: Streamlit launches but shows an ImportError or ModuleNotFoundError in the browser or terminal.
Likely cause: Dependencies were not installed, or were installed at incompatible versions.
Fix: Run pip install -r requirements.txt again in the same Python environment you use to launch the app. Verify versions with pip show tensorflow numpy streamlit.