API Reference

Every function in Pyro's 76 built-in modules.

math

Mathematical functions and constants.

math.PIPi constant (3.14159...)

math.EEuler's number (2.71828...)

math.TAUTau constant (6.28318...)

math.abs(x)Absolute value

math.sqrt(x)Square root

math.pow(x, y)Raise x to the power y

math.ceil(x)Round up to nearest integer

math.floor(x)Round down to nearest integer

math.round(x)Round to nearest integer

math.sin(x)Sine

math.cos(x)Cosine

math.tan(x)Tangent

math.log(x)Natural logarithm

math.log2(x)Base-2 logarithm

math.log10(x)Base-10 logarithm

math.exp(x)e raised to the power x

math.random()Random double in [0, 1)

math.randint(a, b)Random integer in [a, b]

io

File I/O operations.

io.read(path)Read entire file as string

io.write(path, content)Write string to file

io.append(path, content)Append string to file

io.readlines(path)Read file as list of lines

io.exists(path)Check if file exists

io.remove(path)Delete a file

io.mkdir(path)Create directory (recursive)

io.listdir(path)List directory entries

json

JSON parsing and serialization.

json.parse(str)Parse JSON string into JsonValue

json.stringify(value)Convert JsonValue to compact JSON string

json.pretty(value)Convert JsonValue to pretty-printed JSON

time

Time, dates, and scheduling.

time.now()Current time in milliseconds since epoch

time.millis()Alias for now()

time.seconds()Current time in seconds

time.format(ms, fmt)Format timestamp with strftime pattern

time.today()Today's date as YYYY-MM-DD

time.timestamp()Current datetime as YYYY-MM-DD HH:MM:SS

time.year()Current year

time.month()Current month (1-12)

time.day()Current day of month

time.hour()Current hour (0-23)

time.minute()Current minute

time.second()Current second

time.sleep(ms)Sleep for ms milliseconds

time.wait(ms)Alias for sleep()

time.date()Current datetime string

time.timer()Create a Timer with elapsed() and reset()

time.every(interval_ms, func, count)Run func repeatedly at interval

time.after(delay_ms, func)Run func after delay

os

Operating system utilities.

os.env(name)Get environment variable

os.setenv(name, val)Set environment variable

os.platform()Returns "linux", "macos", or "windows"

os.pid()Current process ID

os.cwd()Current working directory

os.cpus()Number of CPU cores

env

Environment variables and .env file loading.

env.get(name, default)Get env var with optional default

env.set(name, val)Set environment variable

env.unset(name)Remove environment variable

env.has(name)Check if env var exists

env.load(path)Load variables from .env file

re

Regular expressions.

re.match(pattern, str)Test if entire string matches pattern

re.has_match(pattern, str)Test if pattern found anywhere in string

re.search(pattern, str)Return first match or empty string

re.replace(pattern, replacement, text)Replace all matches

re.findall(pattern, str)Return all matches as list

re.find_all(pattern, text)Alias for findall()

re.split(pattern, str)Split string by regex pattern

fs

Filesystem operations.

fs.read(path)Read file contents

fs.write(path, data)Write data to file

fs.append(path, data)Append data to file

fs.copy(src, dst)Copy file or directory

fs.move(src, dst)Move/rename file or directory

fs.remove(path)Delete file or directory recursively

fs.mkdir(path)Create directories recursively

fs.list(path)List directory entries

fs.size(path)Get file size in bytes

fs.exists(path)Check if path exists

fs.is_file(path)Check if path is a regular file

fs.is_dir(path)Check if path is a directory

web

Web server framework with routing, middleware, and SSE.

web.app()Create a new web application

app.get(path, handler)Register GET route

app.post(path, handler)Register POST route

app.put(path, handler)Register PUT route

app.del(path, handler)Register DELETE route

app.use(middleware)Add middleware function

app.serve_static(prefix, dir)Serve static files from directory

app.listen(port)Start server on port

app.sse(path, generator, interval)Server-Sent Events endpoint

app.live(path, handler)Long-polling endpoint

web.html(body)Create HTML response

web.json(body)Create JSON response

web.text(body)Create text response

web.redirect(url)Create redirect response

web.bus()Create in-memory message bus

http

HTTP client (powered by libcurl).

http.get(url)HTTP GET request

http.post(url, body)HTTP POST request

http.put(url, body)HTTP PUT request

http.del(url)HTTP DELETE request

http.patch(url, body)HTTP PATCH request

http.request(method, url, body, headers)Custom HTTP request

http.download(url, path)Download file from URL

db

SQLite database (real SQLite via sqlite3).

db.connect(path)Open SQLite database connection

conn.exec(sql)Execute SQL statement

conn.query(sql)Query and return list of Row objects

conn.close()Close database connection

row.get(key)Get column value from Row by name

ai

AI/LLM integration (multi-provider, OpenAI-compatible). You must configure a provider before use.

ai.provider(name, key, model)Configure provider (nvidia, openai, gemini, groq, ollama, etc.) — required before any AI call

ai.chat(prompt, model)Send prompt to AI, get response

ai.chat(prompt, system, model)Chat with system prompt

ai.ask(question)Simple alias for chat()

ai.summarize(text, max_words)Summarize text

ai.translate(text, to)Translate text to target language

ai.classify(text, labels)Classify text into one of the labels

ai.extract(text, fields)Extract structured fields from text as JSON

ai.generate_code(description)Generate Pyro code from description

ai.conversation(system)Create multi-turn conversation

ai.set_key(key)Set API key

ai.set_model(model)Set default model

ai.set_url(url)Set API endpoint URL

ml

Machine learning: regression, classification, clustering.

ml.load_csv(path, target)Load CSV into Dataset

ml.split(dataset, ratio)Split dataset into train/test

ml.normalize(dataset)Min-max normalize features

ml.linear_regression(ds, lr, epochs)Train linear regression model

ml.logistic(ds, lr, epochs)Train logistic regression model

ml.knn_predict(train, test, k)K-nearest neighbors prediction

ml.kmeans(ds, k, max_iter)K-means clustering

ml.mse(pred, actual)Mean squared error

ml.mae(pred, actual)Mean absolute error

ml.r2_score(pred, actual)R-squared score

ml.accuracy(pred, actual)Classification accuracy

nn

Neural networks with backpropagation.

nn.sequential()Create sequential model

nn.dense(in, out, activation)Dense layer (relu, sigmoid, tanh, softmax)

nn.dropout(rate)Dropout regularization layer

nn.randn(rows, cols)Random tensor (He initialization)

model.add(layer)Add layer to model

model.compile(loss, lr)Set loss function and learning rate

model.fit(x, y, in_size, out_size, epochs, batch)Train the model

model.predict(input)Run inference

model.evaluate(x, y, out_size)Evaluate accuracy

model.summary()Print model architecture

model.save(path)Save model weights to file

nn.mse_loss(pred, target)Mean squared error loss

nn.cross_entropy_loss(pred, target)Cross-entropy loss

tensor

N-dimensional tensor operations.

tensor.zeros(r, c)Tensor of zeros

tensor.ones(r, c)Tensor of ones

tensor.eye(n)Identity matrix

tensor.random(r, c)Random tensor in [0, 1)

tensor.range(start, end, step)Tensor from range

tensor.sum(t)Sum of all elements

tensor.mean(t)Mean of all elements

tensor.min_val(t)Minimum element

tensor.max_val(t)Maximum element

tensor.std_dev(t)Standard deviation

tensor.sqrt(t)Element-wise square root

tensor.abs(t)Element-wise absolute value

tensor.exp(t)Element-wise exponential

tensor.log(t)Element-wise natural log

tensor.pow(t, p)Element-wise power

tensor.dot(a, b)Dot product

tensor.matmul(a, b)Matrix multiplication

tensor.transpose(t)Transpose matrix

tensor.det(t)Determinant (2x2 and 3x3)

tensor.inverse(t)Matrix inverse (2x2)

nlp

Natural language processing.

nlp.tokenize(text)Tokenize into words and punctuation

nlp.word_tokenize(text)Tokenize into alphabetic words only

nlp.sent_tokenize(text)Split text into sentences

nlp.lowercase(str)Convert to lowercase

nlp.remove_stopwords(text)Remove common stopwords

nlp.remove_punctuation(text)Strip punctuation characters

nlp.stem(word)Porter-like suffix stemmer

nlp.ngrams(text, n)Generate n-grams

nlp.word_count(text)Word frequency map

nlp.keywords(text, top)Extract top keywords by TF score

nlp.similarity(a, b)Cosine similarity between two texts

nlp.sentiment(text)Lexicon-based sentiment analysis

nlp.ner(text)Rule-based named entity recognition

cv

Computer vision and image processing.

cv.create(w, h, r, g, b)Create blank image

cv.load(path)Load PPM/PGM image

cv.save(img, path)Save image as PPM or BMP

cv.grayscale(img)Convert to grayscale

cv.resize(img, w, h)Resize image (nearest neighbor)

cv.flip(img, dir)Flip horizontal or vertical

cv.blur(img, radius)Box blur

cv.edges(img)Sobel edge detection

cv.threshold(img, thresh)Binary threshold

cv.brightness(img, delta)Adjust brightness

cv.invert(img)Invert colors

cv.crop(img, x, y, w, h)Crop region

cv.draw_rect(img, x, y, w, h, r, g, b)Draw rectangle outline

cv.fill_rect(img, x, y, w, h, r, g, b)Draw filled rectangle

cv.histogram(img)Pixel value distribution (256 bins)

viz

Rich SVG visualizations with dark theme and animation.

viz.bar_chart(labels, values, title, w, h)Animated bar chart

viz.line_chart(x, y, title, w, h)Line chart with gradient fill

viz.scatter(x, y, title, w, h)Animated scatter plot

viz.pie_chart(labels, values, title, w, h)Donut-style pie chart

viz.save(path, svg)Save SVG string to file

plot

Simple SVG chart generation.

plot.set_size(w, h)Set chart dimensions

plot.line(y, title, file)Line chart from y values

plot.bar(labels, values, title, file)Bar chart

plot.scatter(x, y, title, file)Scatter plot

plot.histogram(data, bins, title, file)Histogram

plot.heatmap(matrix, title, file)Heatmap from 2D matrix

plot.training_curve(loss, title, file)Training loss curve

plot.save(file)Save last chart to file

plot.show()Open last chart in browser

test

Unit testing and benchmarking.

test.run(name, fn)Run a named test

test.eq(got, expected)Assert equal

test.neq(got, expected)Assert not equal

test.ok(condition)Assert true

test.fail(msg)Force test failure

test.gt(a, b)Assert a > b

test.lt(a, b)Assert a < b

test.gte(a, b)Assert a >= b

test.lte(a, b)Assert a <= b

test.summary()Print test results summary

test.bench(name, iterations, fn)Benchmark a function

validate

Input validation and sanitization.

validate.email(str)Validate email address

validate.url(str)Validate URL (http/https)

validate.ip(str)Validate IPv4 address

validate.phone(str)Validate phone number

validate.number(str)Check if string is numeric

validate.alpha(str)Check if string is alphabetic

validate.alphanumeric(str)Check if string is alphanumeric

validate.empty(str)Check if string is empty

validate.not_empty(str)Check if string is not empty

validate.length(str, min, max)Check string length in range

validate.sanitize(str)HTML-escape special characters

validate.sql_safe(str)Escape SQL injection characters

validate.password_strength(str)Score password strength (0-6)

crypto

Cryptography via OpenSSL.

crypto.sha256(input)SHA-256 hash

crypto.sha512(input)SHA-512 hash

crypto.hmac_sha256(key, data)HMAC-SHA256

crypto.encrypt(plaintext, key)AES-256-GCM encryption

crypto.decrypt(ciphertext, key)AES-256-GCM decryption

crypto.random_token(length)Secure random alphanumeric token

crypto.random_bytes(length)Secure random bytes (hex)

crypto.uuid()Generate UUID v4

crypto.hash_password(password, iterations)PBKDF2 password hash

crypto.verify_password(password, hash)Verify password against hash

crypto.to_hex(data)Bytes to hex string

crypto.from_hex(hex)Hex string to bytes

csv

CSV reading and writing.

csv.read(path)Read CSV file into 2D string array

csv.parse(str)Parse CSV string

csv.stringify(data)Convert 2D array to CSV string

csv.write(path, data)Write 2D array to CSV file

data

DataFrame for tabular data analysis.

data.read(path)Load CSV into DataFrame

data.write(df, path)Save DataFrame to CSV

df.rows()Number of rows

df.cols()Number of columns

df.col(name)Get column as Series

df.head(n)Print first n rows

df.describe()Print column statistics

df.where(col, predicate)Filter rows

df.sort_by(col, ascending)Sort by column

df.select(cols)Select columns

df.groupby(key, agg_col, func)Group by column and aggregate

df.merge(other, on)Join two DataFrames

df.apply(col, func)Apply function to column

df.drop_duplicates()Remove duplicate rows

df.fillna(val)Fill empty cells

df.sample(n)Random sample of n rows

df.to_json()Convert to JSON string

df.rename(old, new)Rename column

df.drop(col)Drop column

random

Random number generation.

random.random()Random double in [0, 1)

random.randint(a, b)Random integer in [a, b]

random.uniform(a, b)Random double in [a, b)

random.seed(s)Seed the random generator

random.choice(list)Pick random element from list

random.sample(list, n)Pick n random elements

random.shuffle(list)Shuffle list in place

base64

Base64 encoding and decoding.

base64.encode(str)Encode string to Base64

base64.decode(str)Decode Base64 to string

uuid

UUID generation.

uuid.v4()Generate random UUID v4

uuid.nil_uuid()Return nil UUID (all zeros)

uuid.from_string(str)UUID from string (passthrough)

color

Terminal color formatting (ANSI escape codes).

color.red(str)Red text

color.green(str)Green text

color.yellow(str)Yellow text

color.blue(str)Blue text

color.magenta(str)Magenta text

color.cyan(str)Cyan text

color.bold(str)Bold text

color.dim(str)Dim text

color.underline(str)Underlined text

color.bg_red(str)Red background

color.bg_green(str)Green background

log

Leveled logging with timestamps.

log.set_level(level)Set minimum log level (trace/debug/info/warn/error/fatal)

log.trace(...)Log at TRACE level

log.debug(...)Log at DEBUG level

log.info(...)Log at INFO level

log.warn(...)Log at WARN level

log.error(...)Log at ERROR level

log.fatal(...)Log at FATAL level

path

Filesystem path manipulation.

path.join(a, b)Join two path segments

path.dirname(p)Parent directory

path.basename(p)File name component

path.extension(p)File extension

path.exists(p)Check if path exists

path.is_file(p)Check if path is a file

path.is_dir(p)Check if path is a directory

path.absolute(p)Get absolute path