simpyQ

CLI tool for querying simulation CSV data using natural language.

Analyze and visualize simulation data from CSV files using natural language queries interpreted via spaCy and a rule-enhanced NLP pipeline.

---

Features

• Terminal-based CLI tool with live interactive mode
• Natural language processing of signal queries (e.g., mean of Vout)
• Signal name indexing and pretty terminal display using Rich
• Plotting of signal vs time with auto-saving to file
• Logging of results and errors to timestamped log files
• Lightweight, extensible, and easy to adapt to other domains

---

Installation

Note : you need to have Spacy installed on your machine before installing.

🧠 spaCy Setup (Required Before Using simpyq)

Before installing and using the simpyq CLI tool, you need to install spaCy along with its transformer-based English language model.

1. Install spaCy (choose CPU or CUDA based on your system):

For CPU-only systems:

pip install -U pip setuptools wheel
pip install -U 'spacy[cuda11x]'
python -m spacy download en_core_web_trf



```bash
# Create environment (recommended)
conda create -n simpyq_env python=3.10
conda activate simpyq_env

# Install dependencies
pip install -r requirements.txt

# Or install manually
pip install pandas numpy matplotlib rich pyfiglet spacy
python -m spacy download en_core_web_trf

or if you have Spacy already installed then : pip install simpyq

---

Usage

# Run in one-shot mode (process query then exit)
python simpyq/cli.py path/to/data.csv "mean of Vout"

# Show available signals
python simpyq/cli.py path/to/data.csv --show

# Plot signal(s)
python simpyq/cli.py path/to/data.csv --plot Vout Iload --start 0.1 --end 0.9

Interactive Mode

# Start interactive REPL
python simpyq/cli.py path/to/data.csv

Then just type queries like:

mean of Vout
rms of Iin
(mean of Vout) + (rms of Iload) / 2
exit()

---

Example Queries

Query	Description
mean of Vout	Computes the average value of Vout
rms of Iload	Root-mean-square of Iload
max of Temp - min of Temp	Temperature swing
integral of Pin	Energy under the Pin curve
rms of Vout + mean of Iin * 2	Composite math query
plot of Vout and Iin	(Use --plot flag instead)

---

Supported Operations

OPERATIONS = {
    "mean": np.mean,
    "average": np.mean,
    "rms": lambda x: np.sqrt(np.mean(np.square(x))),
    "std": np.std,
    "variance": np.var,
    "max": np.max,
    "min": np.min,
    "abs max": lambda x: np.max(np.abs(x)),
    "abs min": lambda x: np.min(np.abs(x)),
    "sum": np.sum,
    "peak-to-peak": lambda x: np.ptp(x),
    "median": np.median,
    "integral": lambda x: np.trapz(x),
    "squared mean": lambda x: np.mean(np.square(x)),
    "derivative": lambda x: np.gradient(x),
    "diff": np.diff,
}

---

How It Works

• Loads CSV data and extracts signal names
• Injects them as named entities into the spaCy NLP pipeline
• Parses user query and extracts operations and signals
• Computes the operations and evaluates full expressions
• Plots and saves results as needed

---

NLP Training / Fine-tuning

You can customize the NLP model to better understand domain-specific expressions.

To train or adapt:

1. Edit get_nlp() to support additional patterns
2. Add EntityRuler rules:

patterns = [{"label": "ELECTRONIC_SIGNAL", "pattern": name} for name in signal_names]
ruler.add_patterns(patterns)

3. (Optional) Fine-tune spaCy model with labeled data:

python -m spacy train config.cfg --output ./model --paths.train ./train.spacy --paths.dev ./dev.spacy

---

License

GPL v3 License

---

Acknowledgments

Thanks to the open-source community and tools like:

• spaCy
• NumPy
• Matplotlib
• Rich