### Install polars_talib

Source: https://github.com/yvictor/polars_ta_extension/blob/master/README.md

Installs the polars_talib package using pip. This is the primary dependency for using the TA-Lib extension with Polars.

```bash
pip install polars_talib
```

--------------------------------

### Install talib Crate using Cargo

Source: https://github.com/yvictor/polars_ta_extension/blob/master/talib/README.md

This command adds the 'talib' crate as a dependency to your Rust project. Ensure you have Cargo installed and configured.

```bash
cargo add talib
```

--------------------------------

### Python Calculation Example

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/basic.ipynb

This Python code snippet demonstrates a simple arithmetic calculation used in the performance comparison. It involves addition, multiplication, and division.

```python
(1.2 + 19.2) * 1000 / 135
```

--------------------------------

### Calculate Exponential Moving Average (EMA) in Polars

Source: https://context7.com/yvictor/polars_ta_extension/llms.txt

Demonstrates calculating EMA with different time periods using the method chaining syntax (.ta.ema()). The example uses sample closing price data and highlights that earlier values will be null until sufficient data is present.

```python
import polars as pl
import polars_talib as plta

df = pl.DataFrame({
    "close": [100.0, 102.0, 104.0, 103.0, 105.0, 107.0, 106.0, 108.0]
})

# Calculate multiple EMAs
result = df.with_columns(
    pl.col("close").ta.ema(5).alias("ema_5"),
    pl.col("close").ta.ema(10).alias("ema_10"),
    pl.col("close").ta.ema(20).alias("ema_20")
)

print(result)
```

--------------------------------

### Get List of Supported TA-Lib Functions in polars_talib

Source: https://github.com/yvictor/polars_ta_extension/blob/master/README.md

Retrieves a list of all TA-Lib indicator functions that are supported and implemented within the polars_talib extension. This function is useful for discovering available technical analysis tools.

```python
import polars_talib as plta

# list of functions
plta.get_functions()
```

--------------------------------

### Calculate Simple Moving Average (SMA) in Polars

Source: https://context7.com/yvictor/polars_ta_extension/llms.txt

Shows how to calculate SMA using both method chaining (.ta.sma()) and functional syntax (plta.sma()). It includes examples with explicit time periods and default parameters, demonstrating usage with sample OHLCV data.

```python
import polars as pl
import polars_talib as plta

# Sample OHLCV data
df = pl.DataFrame({
    "open": [100.0, 102.0, 101.0, 103.0, 105.0],
    "high": [103.0, 104.0, 103.0, 106.0, 107.0],
    "low": [99.0, 101.0, 100.0, 102.0, 104.0],
    "close": [102.0, 103.0, 101.0, 105.0, 106.0],
    "volume": [1000, 1100, 900, 1200, 1300]
})

# Method syntax: using .ta namespace
result = df.with_columns(
    pl.col("close").ta.sma(timeperiod=3).alias("sma_3")
)

# Functional syntax: using plta module
result = df.with_columns(
    plta.sma(pl.col("close"), timeperiod=3).alias("sma_3")
)

# Default parameters (uses "close" column, timeperiod=30)
result = df.with_columns(
    plta.sma().alias("sma_30")
)

print(result)
```

--------------------------------

### Get TA-Lib Functions Grouped by Category in polars_talib

Source: https://github.com/yvictor/polars_ta_extension/blob/master/README.md

Retrieves a dictionary where TA-Lib indicator functions are grouped by their respective categories (e.g., Overlap Studies, Momentum Indicators). This helps in understanding the organization of available technical indicators.

```python
import polars_talib as plta

# dict of functions by group
plta.get_function_groups()
```

--------------------------------

### Initialize and Explore Polars TA-Lib

Source: https://context7.com/yvictor/polars_ta_extension/llms.txt

Demonstrates importing the library, checking the TA-Lib version, listing all available functions, and retrieving functions grouped by category. This helps in understanding the library's capabilities and structure.

```python
import polars as pl
import polars_talib as plta

# Library initialization happens automatically on import
# Check TA-Lib version
print(plta.__talib_version__)  # "0.4.0"

# List all available functions
functions = plta.get_functions()
print(functions[:5])  # ['ht_dcperiod', 'ht_dcphase', 'ht_phasor', 'ht_sine', 'ht_trendmode']

# Get functions grouped by category
groups = plta.get_function_groups()
print(groups.keys())  # dict_keys(['Cycle Indicators', 'Math Operators', 'Momentum Indicators', ...])
```

--------------------------------

### Import Libraries for Polars and TA

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/bench.ipynb

Imports necessary libraries including polars, polars_talib, pandas, and talib.abstract for technical analysis.

```python
import time
import polars as pl
import polars_talib as plta
import pandas as pd
import talib.abstract as ta
```

--------------------------------

### Calculate WCLPrice with Polars TA

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/bench.ipynb

Calculates the Weighted Closing Price using polars_talib, grouped by 'Symbol', and measures the execution time.

```python
start_t = time.time()
p.with_columns(
    plta.wclprice().over("Symbol").alias("wclprice")
).collect()
end_t = time.time()
pl_wclprice_spend_t = end_t - start_t
spend_records["pl_wclprice"] = pl_wclprice_spend_t
```

--------------------------------

### Calculate WCLPrice with Pandas TA

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/bench.ipynb

Calculates the Weighted Closing Price using pandas and talib.abstract, grouped by 'Ticker', and measures the execution time.

```python
start_t = time.time()
df["wclprice"] = df.groupby("Ticker").apply(lambda x: ta.WCLPRICE(x)).droplevel(0)
end_t = time.time()
pd_wclprice_spend_t = end_t - start_t
spend_records["pd_wclprice"] = pd_wclprice_spend_t
```

--------------------------------

### Read and Prepare Data with Pandas

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/bench.ipynb

Reads a Parquet file into a Pandas DataFrame, sets a multi-index, and converts specific column names to lowercase.

```python
start_t = time.time()
df = (
    pd.read_parquet("us_market_cap2000.parquet")
    .set_index(["Ticker", "Date"])
    .rename(columns={c: c.lower() for c in ["Open", "High", "Low", "Close"]})
)
end_t = time.time()
pd_read_spend = end_t - start_t
spend_records["pd_read"] = pd_read_spend
```

--------------------------------

### Performance Comparison: Polars with polars_talib vs. Pandas with ta

Source: https://github.com/yvictor/polars_ta_extension/blob/master/README.md

Measures the execution time of calculating multiple technical indicators (SMA, MACD, STOCH, WCLPRICE) using both Polars with polars_talib and Pandas with the standard TA-Lib library. This benchmark highlights the significant performance advantage of the Polars-based solution.

```python
%%timeit
df = p.with_columns(
    plta.sma(timeperiod=5).over("Symbol").alias("sma5"),
    plta.macd(fastperiod=10, slowperiod=20, signalperiod=5).over("Symbol").alias("macd"),
    plta.stoch(pl.col("high"), pl.col("low"), pl.col("close"), fastk_period=14, slowk_period=7, slowd_period=7).over("Symbol").alias("stoch"),
    plta.wclprice().over("Symbol").alias("wclprice"),
).with_columns(
    pl.col("macd").struct.field("macd"),
    pl.col("macd").struct.field("macdsignal"),
    pl.col("macd").struct.field("macdhist"),
    pl.col("stoch").struct.field("slowk"),
    pl.col("stoch").struct.field("slowd"),
).select(
    pl.exclude("stoch")
).filter(
    pl.col("Symbol") == "AAPL"
).collect()
```

--------------------------------

### Create Sample OHLCV DataFrame for Trading Strategy in Polars

Source: https://context7.com/yvictor/polars_ta_extension/llms.txt

This Python code snippet demonstrates how to create a sample Polars DataFrame containing typical OHLCV (Open, High, Low, Close, Volume) data with timestamps and symbols, suitable for backtesting trading strategies. It uses Polars' datetime functionality to generate a date range.

```python
import polars as pl
import polars_talib as plta

# Load or create OHLCV data
df = pl.DataFrame({
    "timestamp": pl.datetime_range(pl.datetime(2024, 1, 1), pl.datetime(2024, 3, 1), "1d", eager=True)[:50],
    "symbol": ["AAPL"] * 50,
    "open": [150.0 + i * 0.5 for i in range(50)],
    "high": [152.0 + i * 0.5 for i in range(50)],
    "low": [149.0 + i * 0.5 for i in range(50)],
    "close": [151.0 + i * 0.5 for i in range(50)],
    "volume": [1000000 + i * 10000 for i in range(50)]
})
```

--------------------------------

### Read and Prepare Data with Polars

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/bench.ipynb

Scans a Parquet file into a Polars DataFrame, selecting specific columns and converting float column names to lowercase.

```python
p = pl.scan_parquet("us_market_cap2000.parquet").select(
    pl.col("Date"), pl.col("Ticker").alias("Symbol"), pl.selectors.float().name.to_lowercase()
)
```

--------------------------------

### Create Benchmark DataFrame

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/bench.ipynb

Constructs a Polars DataFrame from the collected performance records, categorizing operations by 'kind' (polars/pandas) and 'op' (read/sma/macd/stoch/wclprice).

```python
df_bench = pl.DataFrame(
    {
        "kind": [k.split("_")[0] for k in spend_records.keys()],
        "op": [k.split("_")[1] for k in spend_records.keys()],
        "time": [v for v in spend_records.values()],
    }
).with_columns(
    pl.when(pl.col("kind")=="pl").then(pl.lit("polars")).otherwise(pl.lit("pandas")).alias("stack")
)
```

--------------------------------

### Performance Comparison: Polars TA-Lib vs. Pandas TA-Lib

Source: https://context7.com/yvictor/polars_ta_extension/llms.txt

This snippet compares the performance of calculating technical indicators (SMA, MACD, RSI) on a large dataset with multiple symbols using Polars with polars_talib extension versus Pandas with the TA-Lib library. It measures the execution time for each approach and calculates the speedup factor.

```python
import polars as pl
import polars_talib as plta
import pandas as pd
import talib
import time

# Create large dataset with multiple symbols
n_symbols = 100
n_bars = 1000

data = []
for symbol_id in range(n_symbols):
    for bar in range(n_bars):
        data.append({
            "Symbol": f"SYM{symbol_id:03d}",
            "close": 100.0 + bar * 0.1,
            "high": 102.0 + bar * 0.1,
            "low": 99.0 + bar * 0.1,
            "open": 100.5 + bar * 0.1
        })

# Polars approach
pl_df = pl.DataFrame(data)
start = time.time()
pl_result = pl_df.with_columns([
    plta.sma(pl.col("close"), timeperiod=5).over("Symbol").alias("sma5"),
    plta.macd(pl.col("close"), fastperiod=10, slowperiod=20, signalperiod=5)
        .over("Symbol")
        .struct.field("macd")
        .alias("macd"),
    plta.rsi(pl.col("close"), timeperiod=14).over("Symbol").alias("rsi")
])
polars_time = time.time() - start
print(f"Polars + polars_talib: {polars_time:.3f}s")

# Pandas approach
pd_df = pd.DataFrame(data)
start = time.time()
pd_df["sma5"] = pd_df.groupby("Symbol")["close"].transform(lambda x: talib.SMA(x, timeperiod=5))
pd_df["macd"] = pd_df.groupby("Symbol")["close"].transform(lambda x: talib.MACD(x, fastperiod=10, slowperiod=20, signalperiod=5)[0])
pd_df["rsi"] = pd_df.groupby("Symbol")["close"].transform(lambda x: talib.RSI(x, timeperiod=14))
pandas_time = time.time() - start
print(f"Pandas + talib: {pandas_time:.3f}s")

print(f"Speedup: {pandas_time / polars_time:.1f}x")
# Expected output: 50-150x faster depending on dataset size and complexity

```

--------------------------------

### Calculate RSI using talib in Rust

Source: https://github.com/yvictor/polars_ta_extension/blob/master/talib/README.md

This Rust code snippet demonstrates how to use the 'talib' crate to calculate the Relative Strength Index (RSI). It includes initializing the library, defining input data (close prices), setting parameters for RSI calculation, processing the results, and shutting down the library. Error handling for the RSI calculation is also included.

```rust
use talib::common::{TimePeriodKwargs, ta_initialize, ta_shutdown};
use talib::momentum::ta_rsi;

fn main() {
    // Initialize Ta-Lib
    let _ = ta_initialize();

    // Sample close prices
    let close_prices: Vec<f64> = vec![
        1.087010, 1.087120, 1.087080, 1.087170, 1.087110, 1.087010, 1.087100, 1.087120, 1.087110,
        1.087080, 1.087000, 1.086630, 1.086630, 1.086610, 1.086630, 1.086640, 1.086650, 1.086650,
        1.086670, 1.086630,
    ];

    // Specify the RSI calculation parameters
    let kwargs = TimePeriodKwargs { timeperiod: 5 };
    // or with builder
    let kwargs = TimePeriodKwargsBuilder::default()
        .timeperiod(5)
        .build()
        .unwrap();

    // Calculate RSI
    let res = ta_rsi(close_prices.as_ptr(), close_prices.len(), &kwargs);

    // Process the result
    match res {
        Ok(rsi_values) => {
            for (index, value) in rsi_values.iter().enumerate() {
                println!("RSI at index {}: {}", index, value);
            }
        }
        Err(e) => {
            println!("Error: {:?}", e);
        }
    }

    // Shutdown Ta-Lib
    let _ = ta_shutdown();
}
```

--------------------------------

### Import polars and polars_talib

Source: https://github.com/yvictor/polars_ta_extension/blob/master/README.md

Imports the necessary libraries, polars for DataFrame manipulation and polars_talib for technical analysis functions.

```python
import polars
import polars_talib as plta
```

--------------------------------

### Download Stock Data with yfinance

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/basic.ipynb

Downloads historical stock data for a list of tickers using the yfinance library. It supports downloading data for up to 2000 tickers and specifies a date range. The data is grouped by ticker and then stacked into a single DataFrame.

```python
import yfinance as yf
import polars as pl
import polars_talib as plta
import pandas as pd
import talib.abstract as ta

tickers = pl.scan_csv("nasdaq_screener.csv").sort("Market Cap", descending=True).filter(
    pl.col("Market Cap").is_not_null() & (pl.col("Market Cap") > 0)
).select(pl.col("Symbol").str.strip_chars(" "), pl.col("Market Cap")).collect()["Symbol"].to_list()

data = yf.download(tickers[:2000], start='2001-01-01', end='2024-12-31', group_by='ticker', threads=32).stack(level=0).reset_index()
```

--------------------------------

### Calculate SMA with Polars TA

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/bench.ipynb

Calculates the Simple Moving Average (SMA) using polars_talib over a 5-period window, grouped by 'Symbol', and measures the execution time.

```python
start_t = time.time()
p.with_columns(
    plta.sma(timeperiod=5).over("Symbol").alias("sma5")
).collect()
end_t = time.time()
pl_sma_spend_t = end_t - start_t
spend_records["pl_sma"] = pl_sma_spend_t
```

--------------------------------

### Calculate Stochastic Oscillator with Polars TA

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/bench.ipynb

Calculates the Stochastic Oscillator (%K and %D) using polars_talib with specified periods, grouped by 'Symbol', and measures the execution time.

```python
start_t = time.time()
p.with_columns(
    plta.stoch(
        pl.col("high"),
        pl.col("low"),
        pl.col("close"),
        fastk_period=14,
        slowk_period=7,
        slowd_period=7,
    )
    .over("Symbol")
    .alias("stoch")
).with_columns(
    pl.col("stoch").struct.field("slowk"),
    pl.col("stoch").struct.field("slowd")
).select(pl.exclude("stoch")).collect()
end_t = time.time()
pl_stoch_spend_t = end_t - start_t
spend_records["pl_stoch"] = pl_stoch_spend_t
```

--------------------------------

### Calculate MACD with Pandas TA

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/bench.ipynb

Calculates the MACD, its signal line, and histogram using pandas and talib.abstract, grouped by 'Ticker', and measures the execution time.

```python
start_t = time.time()
g = df.groupby("Ticker")["close"]
df["macd"] = g.transform(lambda x: ta.MACD(x, fastperiod=10, slowperiod=20, signalperiod=5)[0])
df["macdsignal"] = g.transform(
    lambda x: ta.MACD(x, fastperiod=10, slowperiod=20, signalperiod=5)[1]
)
df["macdhist"] = g.transform(lambda x: ta.MACD(x, fastperiod=10, slowperiod=20, signalperiod=5)[2])
end_t = time.time()
pd_macd_spend_t = end_t - start_t
spend_records["pd_macd"] = pd_macd_spend_t
```

--------------------------------

### Calculate Stochastic Oscillator with Pandas TA

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/bench.ipynb

Calculates the Stochastic Oscillator (%K and %D) using pandas and talib.abstract, grouped by 'Ticker', and measures the execution time.

```python
start_t = time.time()
g = df.groupby("Ticker")
df["slowk"] = g.apply(
    lambda x: ta.STOCH(x, fastk_period=14, slowk_period=7, slowd_period=7)
).droplevel(0)["slowk"]
df["slowd"] = g.apply(
    lambda x: ta.STOCH(x, fastk_period=14, slowk_period=7, slowd_period=7)
).droplevel(0)["slowd"]
end_t = time.time()
pd_stoch_spend_t = end_t - start_t
spend_records["pd_stoch"] = pd_stoch_spend_t
```

--------------------------------

### Use polars_talib Functions Directly like talib.abstract

Source: https://github.com/yvictor/polars_ta_extension/blob/master/README.md

Illustrates how to call TA-Lib functions directly from the plta module, similar to the `talib.abstract` interface. This approach allows for more flexibility in specifying parameters and can be used with or without explicit column selections.

```python
df.with_columns(
    plta.ht_dcperiod(),
    plta.ht_dcperiod(pl.col("close")),
    plta.aroon(),
    plta.aroon(pl.col("high"), pl.col("low"), timeperiod=10),
    plta.wclprice(),
    plta.wclprice(
        pl.col("high"), pl.col("low"), pl.col("close"), 
        timeperiod=10
    ),
)
```

--------------------------------

### Performance Comparison: Pandas with ta

Source: https://github.com/yvictor/polars_ta_extension/blob/master/README.md

Measures the execution time of calculating multiple technical indicators (SMA, MACD, STOCH, WCLPRICE) using Pandas with the standard TA-Lib library. This benchmark is provided for comparison with the Polars-based solution.

```python
%%timeit
df["sma5"] = df.groupby("Ticker")["close"].transform(lambda x: ta.SMA(x, timeperiod=5))
df["macd"] = df.groupby("Ticker")["close"].transform(lambda x: ta.MACD(x, fastperiod=10, slowperiod=20, signalperiod=5)[0])
df["macdsignal"] = df.groupby("Ticker")["close"].transform(lambda x: ta.MACD(x, fastperiod=10, slowperiod=20, signalperiod=5)[1])
df["macdhist"] = df.groupby("Ticker")["close"].transform(lambda x: ta.MACD(x, fastperiod=10, slowperiod=20, signalperiod=5)[2])
df["slowk"] = df.groupby("Ticker").apply(lambda x: ta.STOCH(x, fastk_period=14, slowk_period=7, slowd_period=7)).droplevel(0)["slowk"] 
df["slowd"] = df.groupby("Ticker").apply(lambda x: ta.STOCH(x, fastk_period=14, slowk_period=7, slowd_period=7)).droplevel(0)["slowd"]
df["wclprice"] = df.groupby("Ticker").apply(lambda x: ta.WCLPRICE(x)).droplevel(0)
df.loc["AAPL"]
```

--------------------------------

### Calculate Price Transforms using polars_talib in Polars

Source: https://context7.com/yvictor/polars_ta_extension/llms.txt

This code calculates various price transform indicators (Average Price, Median Price, Typical Price, Weighted Close Price) using the polars_talib library. It demonstrates both explicit column passing and default column name usage. The input is a Polars DataFrame with 'open', 'high', 'low', and 'close' columns.

```python
import polars as pl
import polars_talib as plta

df = pl.DataFrame({
    "open": [100.0, 102.0, 101.0, 103.0, 105.0],
    "high": [103.0, 104.0, 103.0, 106.0, 107.0],
    "low": [99.0, 101.0, 100.0, 102.0, 104.0],
    "close": [102.0, 103.0, 101.0, 105.0, 106.0]
})

# Calculate various price transforms
result = df.with_columns(
    plta.avgprice(pl.col("open"), pl.col("high"), pl.col("low"), pl.col("close"))
        .alias("avg_price"),
    plta.medprice(pl.col("high"), pl.col("low"))
        .alias("median_price"),
    plta.typprice(pl.col("high"), pl.col("low"), pl.col("close"))
        .alias("typical_price"),
    plta.wclprice(pl.col("high"), pl.col("low"), pl.col("close"))
        .alias("weighted_close_price")
)

# Using method syntax with default column names
result = df.with_columns(
    plta.wclprice().alias("wcl")  # Defaults to high/low/close columns
)

print(result)
```

--------------------------------

### Plot Benchmark Results

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/bench.ipynb

Generates a bar chart comparing the execution times ('time') for each operation ('op'), separated by the library stack ('stack').

```python
df_bench.plot.bar(x="op", y="time", by="stack")
```

--------------------------------

### Calculate MACD with Polars TA

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/bench.ipynb

Calculates the MACD indicator using polars_talib with specified periods, extracts MACD, signal line, and histogram, and measures the execution time.

```python
start_t = time.time()
p.with_columns(
    plta.macd(fastperiod=10, slowperiod=20, signalperiod=5).over("Symbol").alias("macd")
).with_columns(
    pl.col("macd").struct.field("macd"),
    pl.col("macd").struct.field("macdsignal"),
    pl.col("macd").struct.field("macdhist")
).collect()
end_t = time.time()
pl_macd_spend_t = end_t - start_t
spend_records["pl_macd"] = pl_macd_spend_t
```

--------------------------------

### Apply SMA using pandas 'apply'

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/basic.ipynb

Calculates SMA for each group (Ticker) using pandas 'apply' method. This method is less efficient than 'transform' and is demonstrated for comparison. The '%%timeit' command measures execution time.

```python
%%timeit
df["sma5"] = df.groupby("Ticker").apply(lambda x: ta.SMA(x, timeperiod=5)).droplevel(0)
```

--------------------------------

### Display Benchmark Results

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/bench.ipynb

Displays the benchmark results DataFrame, showing operation type, time taken, and the library stack (polars or pandas).

```python
df_bench
```

--------------------------------

### Calculate Statistical Functions using polars_talib in Polars

Source: https://context7.com/yvictor/polars_ta_extension/llms.txt

This Python script demonstrates the calculation of various statistical functions including linear regression (slope, intercept, angle), correlation, beta, standard deviation, variance, and time series forecast using the polars_talib library. It requires input DataFrames with appropriate columns (e.g., 'x', 'y') and specified time periods.

```python
import polars as pl
import polars_talib as plta

df = pl.DataFrame({
    "x": [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0],
    "y": [2.0, 4.0, 5.0, 4.0, 5.0, 7.0, 8.0, 9.0, 10.0, 11.0]
})

# Linear regression
result = df.with_columns(
    plta.linearreg(pl.col("x"), timeperiod=5).alias("linreg"),
    plta.linearreg_slope(pl.col("x"), timeperiod=5).alias("slope"),
    plta.linearreg_intercept(pl.col("x"), timeperiod=5).alias("intercept"),
    plta.linearreg_angle(pl.col("x"), timeperiod=5).alias("angle")
)

# Correlation and Beta
result = df.with_columns(
    plta.correl(pl.col("x"), pl.col("y"), timeperiod=5).alias("correlation"),
    plta.beta(pl.col("x"), pl.col("y"), timeperiod=5).alias("beta")
)

# Standard deviation and variance
result = df.with_columns(
    plta.stddev(pl.col("x"), timeperiod=5, nbdev=1.0).alias("stddev"),
    plta.var(pl.col("x"), timeperiod=5, nbdev=1.0).alias("variance")
)

# Time Series Forecast
result = df.with_columns(
    plta.tsf(pl.col("x"), timeperiod=5).alias("forecast")
)

print(result)
```

--------------------------------

### Calculate SMA with Pandas TA

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/bench.ipynb

Calculates the Simple Moving Average (SMA) for a 5-period window on the 'close' price, grouped by 'Ticker', and measures the execution time.

```python
start_t = time.time()
df["sma5"] = df.groupby("Ticker")["close"].transform(lambda x: ta.SMA(x, timeperiod=5))
end_t = time.time()
pd_sma_spend_t = end_t - start_t
spend_records["pd_sma"] = pd_sma_spend_t
```

--------------------------------

### Apply SMA using pandas 'transform'

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/basic.ipynb

Calculates SMA for each group (Ticker) using pandas 'transform' method. This is generally faster than 'apply' but still significantly slower than polars' optimized approach. The '%%timeit' command measures execution time.

```python
%%timeit
df["sma5"] = df.groupby("Ticker")["close"].transform(lambda x: ta.SMA(x, timeperiod=5))
```

--------------------------------

### Apply SMA using polars_talib with 'over' syntax

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/basic.ipynb

Applies the Simple Moving Average (SMA) calculation to each symbol in a DataFrame using the polars_talib extension. The 'over' syntax enables efficient group-wise operations. This method is significantly faster than pandas equivalents, including file reading and data transformation.

```python
%%timeit
df = p.with_columns(
    plta.sma(timeperiod=5).over("Symbol").alias("sma5"),
).filter(
    pl.col("Symbol") == "NVDA"
).collect()
```

--------------------------------

### Benchmark Polars Data Reading

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/bench.ipynb

Measures the time taken to collect data into a Polars DataFrame and records it.

```python
spend_records = {}
start_t = time.time()
p.collect()
end_t = time.time()
pl_read_spend = end_t - start_t
spend_records["pl_read"] = pl_read_spend
```

--------------------------------

### Scan Parquet and Prepare for Polars TA

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/basic.ipynb

Scans a parquet file containing stock data and selects specific columns. It renames the 'Ticker' column to 'Symbol' and converts all float column names to lowercase. This is a preparatory step for using polars_talib functions.

```python
p = pl.scan_parquet("us_market_cap2000.parquet").select(
    pl.col("Date"), pl.col("Ticker").alias("Symbol"),
    pl.selectors.float().name.to_lowercase()
)
```

--------------------------------

### Read and transform data with pandas

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/basic.ipynb

Reads a Parquet file and renames columns to lowercase using pandas. This operation is used as a baseline for performance comparison. The '%%timeit' magic command measures the execution time.

```python
%%timeit
df = pd.read_parquet("us_market_cap2000.parquet").set_index(["Ticker", "Date"]).rename(
    columns={c: c.lower() for c in ["Open", "High", "Low", "Close"]}
)
```

```python
df = pd.read_parquet("us_market_cap2000.parquet").set_index(["Ticker", "Date"]).rename(
    columns={c: c.lower() for c in ["Open", "High", "Low", "Close"]}
)
```

--------------------------------

### Apply multiple TA functions using pandas

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/basic.ipynb

Applies multiple technical analysis functions (SMA, MACD, STOCH, WCLPRICE) using pandas 'transform' and 'apply'. This demonstrates the more verbose and less efficient syntax in pandas, especially for functions with multiple outputs, leading to significantly longer execution times.

```python
%%timeit
df["sma5"] = df.groupby("Ticker")["close"].transform(lambda x: ta.SMA(x, timeperiod=5))
df["macd"] = df.groupby("Ticker")["close"].transform(lambda x: ta.MACD(x, fastperiod=10, slowperiod=20, signalperiod=5)[0])
df["macdsignal"] = df.groupby("Ticker")["close"].transform(lambda x: ta.MACD(x, fastperiod=10, slowperiod=20, signalperiod=5)[1])
df["macdhist"] = df.groupby("Ticker")["close"].transform(lambda x: ta.MACD(x, fastperiod=10, slowperiod=20, signalperiod=5)[2])
df["slowk"] = df.groupby("Ticker").apply(lambda x: ta.STOCH(x, fastk_period=14, slowk_period=7, slowd_period=7)).droplevel(0)["slowk"]
df["slowd"] = df.groupby("Ticker").apply(lambda x: ta.STOCH(x, fastk_period=14, slowk_period=7, slowd_period=7)).droplevel(0)["slowd"]
df["wclprice"] = df.groupby("Ticker").apply(lambda x: ta.WCLPRICE(x)).droplevel(0)
```

--------------------------------

### Calculate Technical Indicators on Multi-Symbol DataFrames using 'over' in Polars

Source: https://github.com/yvictor/polars_ta_extension/blob/master/README.md

Shows how to apply TA-Lib indicators to a Polars DataFrame containing multiple symbols using the `.over("symbol")` syntax. This is useful for calculations that should be performed independently for each group (e.g., each stock symbol).

```python
df.with_columns(
    pl.col("close").ta.ema(5).over("symbol").alias("ema5"),
    pl.col("close").ta.macd(12, 26, 9).over("symbol").struct.field("macd"),
    pl.col("close").ta.macd(12, 26, 9).over("symbol").struct.field("macdsignal"),
    pl.col("open").ta.cdl2crows(
        pl.col("high"), pl.col("low"), pl.col("close")
    ).over("symbol").alias("cdl2crows"),
    pl.col("close").ta.wclprice("high", "low").over("symbol").alias("wclprice"),
)
```

--------------------------------

### Calculate Multiple Technical Indicators and Trading Signals in Polars

Source: https://context7.com/yvictor/polars_ta_extension/llms.txt

This snippet demonstrates how to calculate various technical indicators (SMA, EMA, RSI, MACD, ATR, Bollinger Bands, OBV) and generate buy/sell trading signals using the polars_ta_extension library within a Polars DataFrame. It also shows how to extract components from complex indicator outputs and filter signals.

```python
import polars as pl
import polars_talib as plta

# Assuming df is a Polars DataFrame with columns like 'close', 'high', 'low', 'volume', 'timestamp'
# Example DataFrame creation (replace with your actual data loading)
df = pl.DataFrame({
    "timestamp": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20],
    "close": [10, 11, 12, 11, 12, 13, 14, 13, 14, 15, 16, 15, 16, 17, 18, 17, 18, 19, 20, 19],
    "high": [11, 12, 13, 12, 13, 14, 15, 14, 15, 16, 17, 16, 17, 18, 19, 18, 19, 20, 21, 20],
    "low": [9, 10, 11, 10, 11, 12, 13, 12, 13, 14, 15, 14, 15, 16, 17, 16, 17, 18, 19, 18],
    "volume": [100, 110, 120, 115, 125, 130, 140, 135, 145, 150, 160, 155, 165, 170, 180, 175, 185, 190, 200, 195],
})

# Calculate multiple indicators
signals = df.with_columns([
    # Trend indicators
    plta.sma(pl.col("close"), timeperiod=10).alias("sma_10"),
    plta.sma(pl.col("close"), timeperiod=20).alias("sma_20"),
    plta.ema(pl.col("close"), timeperiod=12).alias("ema_12"),

    # Momentum indicators
    plta.rsi(pl.col("close"), timeperiod=14).alias("rsi"),
    plta.macd(pl.col("close"), fastperiod=12, slowperiod=26, signalperiod=9).alias("macd_struct"),

    # Volatility
    plta.atr(pl.col("high"), pl.col("low"), pl.col("close"), timeperiod=14).alias("atr"),
    plta.bbands(pl.col("close"), timeperiod=20, nbdevup=2.0, nbdevdn=2.0).alias("bbands"),

    # Volume
    plta.obv(pl.col("close"), pl.col("volume")).alias("obv"),
]).with_columns([
    # Extract MACD components
    pl.col("macd_struct").struct.field("macd").alias("macd"),
    pl.col("macd_struct").struct.field("macdsignal").alias("macd_signal"),
    pl.col("macd_struct").struct.field("macdhist").alias("macd_hist"),

    # Extract Bollinger Bands
    pl.col("bbands").struct.field("upperband").alias("bb_upper"),
    pl.col("bbands").struct.field("middleband").alias("bb_middle"),
    pl.col("bbands").struct.field("lowerband").alias("bb_lower"),
]).drop(["macd_struct", "bbands"])

# Generate trading signals
signals = signals.with_columns([
    # Bullish signal: SMA crossover + RSI not overbought + MACD positive
    (
        (pl.col("sma_10") > pl.col("sma_20")) &
        (pl.col("sma_10").shift(1) <= pl.col("sma_20").shift(1)) &
        (pl.col("rsi") < 70) &
        (pl.col("macd") > pl.col("macd_signal"))
    ).alias("buy_signal"),

    # Bearish signal: SMA crossunder + RSI not oversold + MACD negative
    (
        (pl.col("sma_10") < pl.col("sma_20")) &
        (pl.col("sma_10").shift(1) >= pl.col("sma_20").shift(1)) &
        (pl.col("rsi") > 30) &
        (pl.col("macd") < pl.col("macd_signal"))
    ).alias("sell_signal"),
])

# Filter to signal rows
buy_signals = signals.filter(pl.col("buy_signal"))
sell_signals = signals.filter(pl.col("sell_signal"))

print("Buy Signals:")
print(buy_signals.select(["timestamp", "close", "rsi", "macd", "sma_10", "sma_20"]))
print("\nSell Signals:")
print(sell_signals.select(["timestamp", "close", "rsi", "macd", "sma_10", "sma_20"]))

```

--------------------------------

### Calculate Hilbert Transform Indicators in Polars

Source: https://context7.com/yvictor/polars_ta_extension/llms.txt

This Python code calculates various Hilbert Transform indicators (Dominant Cycle Period, Trend vs Cycle Mode, Sine Wave, Phasor Components) using the polars_talib library. It handles indicators that return single values, structs, and demonstrates extracting specific fields from struct outputs. The input is a Polars DataFrame with a 'close' column.

```python
import polars as pl
import polars_talib as plta

df = pl.DataFrame({
    "close": [100.0, 102.0, 101.0, 103.0, 105.0, 104.0, 106.0, 108.0, 107.0, 109.0,
              110.0, 112.0, 111.0, 113.0, 115.0, 114.0, 116.0, 118.0, 117.0, 119.0] * 3
})

# Hilbert Transform - Dominant Cycle Period
result = df.with_columns(
    plta.ht_dcperiod(pl.col("close")).alias("ht_dcperiod")
)

# Hilbert Transform - Trend vs Cycle Mode
result = df.with_columns(
    plta.ht_trendmode(pl.col("close")).alias("ht_trendmode")
)

# Hilbert Transform - Sine Wave (returns struct)
result = df.with_columns(
    plta.ht_sine(pl.col("close")).alias("ht_sine_struct")
).with_columns(
    pl.col("ht_sine_struct").struct.field("sine").alias("sine"),
    pl.col("ht_sine_struct").struct.field("leadsine").alias("leadsine")
).drop("ht_sine_struct")

# Hilbert Transform - Phasor Components
result = df.with_columns(
    plta.ht_phasor().alias("phasor")
).with_columns(
    pl.col("phasor").struct.field("inphase").alias("inphase"),
    pl.col("phasor").struct.field("quadrature").alias("quadrature")
).drop("phasor")

print(result)
```

--------------------------------

### Multi-Symbol Analysis with .over() in Polars TA

Source: https://context7.com/yvictor/polars_ta_extension/llms.txt

Demonstrates performing technical analysis calculations on data containing multiple symbols. Although the code snippet itself does not contain TA calculations, it sets up a DataFrame structured for such analysis, implying that subsequent TA functions would be applied using `.over('symbol')` to compute indicators independently for each symbol.

```python
import polars as pl
import polars_talib as plta

# Data with multiple symbols
df = pl.DataFrame({
    "symbol": ["AAPL"] * 10 + ["GOOGL"] * 10 + ["MSFT"] * 10,
    "timestamp": list(range(10)) * 3,
    "close": [100, 102, 101, 103, 105, 104, 106, 108, 107, 109] * 3,
    "high": [103, 104, 103, 106, 107, 106, 108, 110, 109, 111] * 3,
    "low": [99, 101, 100, 102, 104, 103, 105, 107, 106, 108] * 3,
    "volume": [1000, 1100, 900, 1200, 1300, 1150, 1250, 1350, 1400, 1500] * 3
})
```

--------------------------------

### Calculate Stochastic Oscillator using Polars TA Extension

Source: https://context7.com/yvictor/polars_ta_extension/llms.txt

Calculates the Stochastic Oscillator, which includes the %K (fast or slow) and %D lines. This indicator compares a particular closing price of a security to a range of its prices over a certain period. It is used to generate overbought and oversold signals.

```python
import polars as pl
import polars_talib as plta

df = pl.DataFrame({
    "high": [103.0, 104.0, 105.0, 106.0, 107.0, 108.0, 109.0, 110.0, 111.0, 112.0] * 2,
    "low": [99.0, 100.0, 101.0, 102.0, 103.0, 104.0, 105.0, 106.0, 107.0, 108.0] * 2,
    "close": [102.0, 103.0, 104.0, 105.0, 106.0, 107.0, 108.0, 109.0, 110.0, 111.0] * 2
})

# Stochastic returns %K and %D lines in a struct
result = df.with_columns(
    plta.stoch(
        pl.col("high"),
        pl.col("low"),
        pl.col("close"),
        fastk_period=14,
        slowk_period=3,
        slowk_matype=0,
        slowd_period=3,
        slowd_matype=0
    ).alias("stoch")
).with_columns(
    pl.col("stoch").struct.field("slowk").alias("stoch_k"),
    pl.col("stoch").struct.field("slowd").alias("stoch_d")
).drop("stoch")

print(result)
# %K and %D oscillate between 0 and 100
```

--------------------------------

### Calculate Technical Indicators with Polars TA Extension (Python)

Source: https://github.com/yvictor/polars_ta_extension/blob/master/examples/basic.ipynb

This Python code snippet demonstrates how to use the Polars TA extension to calculate multiple technical indicators including SMA, MACD, Stochastic Oscillator, and Weighted Close Price. It applies these indicators over different 'Symbol' groups and then extracts specific fields from the MACD and Stochastic outputs. Finally, it filters the data for a specific symbol ('AAPL') and collects the results.

```python
p.with_columns(
    plta.sma(timeperiod=5).over("Symbol").alias("sma5"),
    plta.macd(fastperiod=10, slowperiod=20, signalperiod=5).over("Symbol").alias("macd"),
    plta.stoch(pl.col("high"), pl.col("low"), pl.col("close"), fastk_period=14, slowk_period=7, slowd_period=7).over("Symbol").alias("stoch"),
    plta.wclprice().over("Symbol").alias("wclprice"),
).with_columns(
    pl.col("macd").struct.field("macd"),
    pl.col("macd").struct.field("macdsignal"),
    pl.col("macd").struct.field("macdhist"),
    pl.col("stoch").struct.field("slowk"),
    pl.col("stoch").struct.field("slowd"),
).select(
    pl.exclude("stoch")
).filter(
    pl.col("Symbol") == "AAPL"
).collect()
```

--------------------------------

### Calculate Technical Indicators on Single Symbol DataFrames in Polars

Source: https://github.com/yvictor/polars_ta_extension/blob/master/README.md

Demonstrates how to apply various TA-Lib indicators (EMA, MACD, CDL2Crows, WCLPRICE) to a Polars DataFrame for a single symbol. It utilizes the `.ta` accessor provided by polars_talib.

```python
df.with_columns(
    pl.col("close").ta.ema(5).alias("ema5"),
    pl.col("close").ta.macd(12, 26, 9).struct.field("macd"),
    pl.col("close").ta.macd(12, 26, 9).struct.field("macdsignal"),
    pl.col("open").ta.cdl2crows(pl.col("high"), pl.col("low"), pl.col("close")).alias("cdl2crows"),
    pl.col("close").ta.wclprice("high", "low").alias("wclprice"),
)
```