nano-wait 3.1.2

Adaptive waiting and smart automation library — includes Wi-Fi, system context, and Vision Mode for screen-based decisions.

Nano-Wait

Intelligent Automation with Adaptive Waiting and Computer Vision

---

Overview

Nano-Wait is a Python library for automating graphical user interfaces (GUIs) that replaces the use of time.sleep() with an Intelligent Adaptive Waiting system, dynamically adjusting the wait time based on:

• Computer performance (CPU and memory)

• Wi-Fi signal quality (when available)

• User-defined aggressiveness level

From version 3.0, Nano-Wait also includes a Computer Vision (OCR) module capable of reading numbers directly from the screen and making automated decisions.

🚀 Why not use time.sleep()?

time.sleep() is static and "blind":

it ignores whether the system is overloaded or if the network is slow.

Nano-Wait solves this by applying a Dynamic Adjustment Factor, ensuring that the script:

• Is not too slow when the system is fast

• Nor too fast to the point of breaking the automation

---

📦 Installation

</code></pre>
<ul>
<li>
<p>Optional Dependencies</p>
</li>
<li>
<p>For full functionality of the Vision module:</p>
</li>
<li>
<p>Tesseract OCR (required for OCR)</p>
</li>
<li>
<p>pytesseract</p>
</li>
<li>
<p>Pillow</p>
</li>
<li>
<p>pyautogui</p>
</li>
<li>
<p>pynput</p>
</li>
<li>
<p>psutil</p>
</li>
<li>
<p>pywifi (Windows only)</p>
</li>
</ul>
<h2>⚠️ Nano-Wait does not collect network data.</h2>
<p>It only reads local signal and operating system performance metrics.</p>
<h2>🧠 Module 1 — Adaptive Waiting (Smart Wait)</h2>
<p>Main function: <strong>wait()</strong></p>
<p>The wait() function is the direct replacement for time.sleep().</p>
<pre><code>from nano_wait import wait

wait(5)

Function Signature

wait(

t: float,
wifi: str | None = None,
speed: str | float = "normal",
verbose: bool = False,
log: bool = False
) -> float

Parameters

Parameter	Default Value	Behavior when omitted
t	required	Defines the maximum wait time. Cannot be omitted.
wifi	None	Nano-Wait ignores network metrics and calculates the factor based only on local performance (CPU and memory).
speed	"normal"	Uses balanced aggressiveness, prioritizing stability without sacrificing performance.
verbose	False	No calculation information is displayed in the terminal.
log	False	No log file is generated (nano_wait.log is neither created nor updated).

Example with Wi-Fi

wait(
5,
wifi="My5G_Network",
speed="fast",
verbose=True

)

Example without Wi-Fi (local hardware only)

wait(2, speed="ultra")

🔬 How wait time is calculated

Nano-Wait calculates an adaptive factor based on:

• CPU usage

• Memory usage

• Wi-Fi signal strength (when available)

Formula applied

wait_time = max(0.05, min(t / factor, t))

Safety rules

• Floor: never wait less than 50 ms

• Ceiling: never exceed the original t time

• Avoids excessive CPU usage

🧠 Module 2 — Vision (OCR and Visual Decision)

The Vision module allows you to read Display numbers on the screen and make automatic decisions.

Main Class

</code></pre>
<h3>Available Modes</h3>
<table>
<thead>
<tr>
<th>Mode</th>
<th>Description</th>
</tr>
</thead>
<tbody>
<tr>
<td>observe</td>
<td>Only reads and displays data</td>
</tr>
<tr>
<td>decision</td>
<td>Reads data and executes actions</td>
</tr>
<tr>
<td>learn</td>
<td>Collects visual patterns (experimental)</td>
</tr>
</tbody>
</table>
<h2>📸 Screen Region Capture</h2>
<p>The user can manually mark regions:</p>
<pre><code>region = VisionMode.mark_region()

The return is a tuple:

(x, y, width, height)

🔍 Complete Example — Reading and Decision

from nano_wait.vision import VisionMode

vision = VisionMode(mode="decision")
region = VisionMode.mark_region()
vision.run(regions=[region])

Internal Logic (decision mode)

If detected number > 1000 → double-click

Otherwise → skip item

These actions can be easily customized in the code.

⚙️ Vision Internal Pipeline

• Screen region capture (ImageGrab)

• Grayscale conversion

• OCR via Tesseract

• Numerical extraction with Regex

• Execution of automated actions

🧪 Learn Mode (Current State)

• The learn mode currently:

• Captures visual data repeatedly

• Serves as a basis for future versions with persistence

📌 Note: The learn mode does not yet save models to disk.

It is experimental and focused on data collection.

System	Wi-Fi	Note
Windows	✅	Uses pywifi
macOS	✅	Uses airport command
Linux	✅	Uses nmcli
Others	❌	Only wireless mode

🛠 Real-World Use Cases

• Visual Automation Bots

• Legacy Dashboard Reading

• Intelligent Click Adjustment

• OCR-Based Automation

• Lightweight RPA without Selenium

🤝 Contribution

• Fork the project

• Create a branch (feature/my-improvement)

• Submit a Pull Request

📄 License

MIT License

👤 Author

Luiz Seabra De Marco

👤 Documentation Author

Vitor Seabra De Marco