PRE

di_spreads(date, bps=False)

Calcula o DI Spread para títulos prefixados (LTN e NTN-F) em uma data de referência.

spread = taxa indicativa do PRE - taxa de ajuste do DI

Quando bps=False a coluna retorna essa diferença em formato decimal (ex: 0.000439 ≈ 4.39 bps). Quando bps=True o valor é automaticamente multiplicado por 10_000 e exibido diretamente em basis points.

Parameters:

Name	Type	Description	Default
date	DateScalar	Data de referência para buscar as taxas.	required
bps	bool	Se True, retorna DISpread já convertido em basis points. Default False.	False

Returns:

Type	Description
DataFrame	pl.DataFrame com colunas: - BondType - MaturityDate - DISpread (decimal ou bps conforme parâmetro)

Examples:

>>> from pyield import pre
>>> pre.di_spreads("30-05-2025", bps=True)
shape: (18, 3)
┌──────────┬──────────────┬──────────┐
│ BondType ┆ MaturityDate ┆ DISpread │
│ ---      ┆ ---          ┆ ---      │
│ str      ┆ date         ┆ f64      │
╞══════════╪══════════════╪══════════╡
│ LTN      ┆ 2025-07-01   ┆ 4.39     │
│ LTN      ┆ 2025-10-01   ┆ -9.0     │
│ LTN      ┆ 2026-01-01   ┆ -4.88    │
│ LTN      ┆ 2026-04-01   ┆ -4.45    │
│ LTN      ┆ 2026-07-01   ┆ 0.81     │
│ …        ┆ …            ┆ …        │
│ NTN-F    ┆ 2027-01-01   ┆ -3.31    │
│ NTN-F    ┆ 2029-01-01   ┆ 14.21    │
│ NTN-F    ┆ 2031-01-01   ┆ 21.61    │
│ NTN-F    ┆ 2033-01-01   ┆ 11.51    │
│ NTN-F    ┆ 2035-01-01   ┆ 22.0     │
└──────────┴──────────────┴──────────┘

Source code in pyield/tn/pre.py

def di_spreads(date: DateScalar, bps: bool = False) -> pl.DataFrame:
    """
    Calcula o DI Spread para títulos prefixados (LTN e NTN-F) em uma data de referência.

    spread = taxa indicativa do PRE - taxa de ajuste do DI

    Quando ``bps=False`` a coluna retorna essa diferença em formato decimal
    (ex: 0.000439 ≈ 4.39 bps). Quando ``bps=True`` o valor é automaticamente
    multiplicado por 10_000 e exibido diretamente em basis points.

    Args:
        date (DateScalar): Data de referência para buscar as taxas.
        bps (bool): Se True, retorna DISpread já convertido em basis points.
            Default False.

    Returns:
        pl.DataFrame com colunas:
            - BondType
            - MaturityDate
            - DISpread (decimal ou bps conforme parâmetro)

    Examples:
        >>> from pyield import pre
        >>> pre.di_spreads("30-05-2025", bps=True)
        shape: (18, 3)
        ┌──────────┬──────────────┬──────────┐
        │ BondType ┆ MaturityDate ┆ DISpread │
        │ ---      ┆ ---          ┆ ---      │
        │ str      ┆ date         ┆ f64      │
        ╞══════════╪══════════════╪══════════╡
        │ LTN      ┆ 2025-07-01   ┆ 4.39     │
        │ LTN      ┆ 2025-10-01   ┆ -9.0     │
        │ LTN      ┆ 2026-01-01   ┆ -4.88    │
        │ LTN      ┆ 2026-04-01   ┆ -4.45    │
        │ LTN      ┆ 2026-07-01   ┆ 0.81     │
        │ …        ┆ …            ┆ …        │
        │ NTN-F    ┆ 2027-01-01   ┆ -3.31    │
        │ NTN-F    ┆ 2029-01-01   ┆ 14.21    │
        │ NTN-F    ┆ 2031-01-01   ┆ 21.61    │
        │ NTN-F    ┆ 2033-01-01   ┆ 11.51    │
        │ NTN-F    ┆ 2035-01-01   ┆ 22.0     │
        └──────────┴──────────────┴──────────┘
    """
    # Fetch bond rates, filtering for LTN and NTN-F types
    df = (
        tpf.tpf_data(date, "PRE")
        .with_columns(DISpread=pl.col("IndicativeRate") - pl.col("DIRate"))
        .select("BondType", "MaturityDate", "DISpread")
        .sort("BondType", "MaturityDate")
    )

    if bps:
        df = df.with_columns(pl.col("DISpread") * 10_000)

    return df

spot_rates(date)

Create the PRE curve (zero coupon rates) for Brazilian fixed rate bonds.

This function combines LTN rates (which are already zero coupon) with spot rates derived from NTN-F bonds using the bootstrap method.

Parameters:

Name	Type	Description	Default
date	DateScalar	The reference date for fetching the data.	required

Returns:

Type	Description
DataFrame	pd.DataFrame: DataFrame with columns "MaturityDate", "BDToMat", and "SpotRate". Contains zero coupon rates for all available maturities.

Raises:

Type	Description
ValueError	If any maturity date cannot be processed or business days cannot be calculated.

Examples:

>>> from pyield import pre
>>> pre.spot_rates("18-06-2025")
   MaturityDate  BDToMat  SpotRate
0    2025-07-01        8   0.14835
1    2025-10-01       74  0.147463
2    2026-01-01      138  0.147752
3    2026-04-01      199  0.147947
4    2026-07-01      260  0.147069
5    2026-10-01      325  0.144733
6    2027-01-01      387  0.142496
7    2027-04-01      447  0.140924
8    2027-07-01      510  0.139024
9    2028-01-01      638  0.136595
10   2028-07-01      762  0.135664
11   2029-01-01      886  0.136484
12   2030-01-01     1135  0.137279
13   2031-01-01     1387  0.138154
14   2032-01-01     1639   0.13876
15   2033-01-01     1891    0.1393
16   2035-01-01     2390  0.141068

Source code in pyield/tn/pre.py

def spot_rates(date: DateScalar) -> pd.DataFrame:
    """
    Create the PRE curve (zero coupon rates) for Brazilian fixed rate bonds.

    This function combines LTN rates (which are already zero coupon) with
    spot rates derived from NTN-F bonds using the bootstrap method.

    Args:
        date (DateScalar): The reference date for fetching the data.

    Returns:
        pd.DataFrame: DataFrame with columns "MaturityDate", "BDToMat", and "SpotRate".
                     Contains zero coupon rates for all available maturities.

    Raises:
        ValueError: If any maturity date cannot be processed or business days cannot be
            calculated.

    Examples:
        >>> from pyield import pre
        >>> pre.spot_rates("18-06-2025")
           MaturityDate  BDToMat  SpotRate
        0    2025-07-01        8   0.14835
        1    2025-10-01       74  0.147463
        2    2026-01-01      138  0.147752
        3    2026-04-01      199  0.147947
        4    2026-07-01      260  0.147069
        5    2026-10-01      325  0.144733
        6    2027-01-01      387  0.142496
        7    2027-04-01      447  0.140924
        8    2027-07-01      510  0.139024
        9    2028-01-01      638  0.136595
        10   2028-07-01      762  0.135664
        11   2029-01-01      886  0.136484
        12   2030-01-01     1135  0.137279
        13   2031-01-01     1387  0.138154
        14   2032-01-01     1639   0.13876
        15   2033-01-01     1891    0.1393
        16   2035-01-01     2390  0.141068
    """
    # Fetch LTN data (zero coupon bonds)
    df_ltn = anbima.tpf_data(date, "LTN").to_pandas(use_pyarrow_extension_array=True)

    # Fetch NTN-F data (coupon bonds)
    df_ntnf = anbima.tpf_data(date, "NTN-F").to_pandas(use_pyarrow_extension_array=True)

    # Check if we have data for both bond types
    if df_ltn.empty and df_ntnf.empty:
        return pd.DataFrame(columns=["MaturityDate", "BDToMat", "SpotRate"])

    # If we only have LTN data, return it directly
    if df_ntnf.empty:
        pass
        # return _process_ltn_only(date, df_ltn)

    # If we only have NTN-F data, we can't bootstrap without LTN rates
    if df_ltn.empty:
        raise ValueError(
            "Cannot construct PRE curve without LTN rates for bootstrapping"
        )

    # Use the existing spot_rates function to calculate zero coupon rates
    df_spots = ntnf.spot_rates(
        settlement=date,
        ltn_maturities=df_ltn["MaturityDate"],
        ltn_rates=df_ltn["IndicativeRate"],
        ntnf_maturities=df_ntnf["MaturityDate"],
        ntnf_rates=df_ntnf["IndicativeRate"],
        show_coupons=False,
    ).to_pandas(use_pyarrow_extension_array=True)

    # Find LTN maturities that are not in the NTN-F result
    ltn_mask = ~df_ltn["MaturityDate"].isin(df_spots["MaturityDate"])
    ltn_not_in_ntnf = df_ltn.loc[ltn_mask].copy()

    if not ltn_not_in_ntnf.empty:
        # Process additional LTN maturities
        ltn_subset = _process_additional_ltn(date, ltn_not_in_ntnf)

        # Ensure consistent data types
        ltn_subset["BDToMat"] = ltn_subset["BDToMat"].astype("int64[pyarrow]")
        df_spots["BDToMat"] = df_spots["BDToMat"].astype("int64[pyarrow]")

        # Combine LTN and NTN-F derived spot rates
        df_combined = pd.concat([df_spots, ltn_subset], ignore_index=True)
    else:
        df_combined = df_spots.copy()
        df_combined["BDToMat"] = df_combined["BDToMat"].astype("int64[pyarrow]")

    # Final validation - ensure no NaN values in the result
    _validate_final_result(df_combined)

    # Sort by maturity date and return
    return df_combined.sort_values("MaturityDate").reset_index(drop=True)