oasislmf.pytools.lec.aggreports.write_tables

Functions

get_loss(next_retperiod, last_retperiod, last_loss, ...)

Get loss based on current and next return period

fill_tvar(tail, tail_sizes, tail_offsets, summary_id, ...)

Populate the Tail with retperiod and tvar values for summary_id

fill_tvar_wheatsheaf(tail, tail_sizes, tail_offsets, ...)

Populate the Tail with retperiod and tvar values for (summary_id, sidx) pair

write_return_period_out(next_returnperiod_idx, ...[, ...])

Processes return periods and computes losses for a given summary, updating TVaR and mean map if required.

write_tvar(epcalc, eptype_tvar, tail, tail_sizes, ...)

Get TVaR values for EPT output from tail

write_tvar_wheatsheaf(num_sidxs, eptype_tvar, tail, ...)

Get TVaR values for PSEPT output from tail

write_ept(buffer, items, items_start_end, ...[, ...])

Generate Loss Exceedance Curve values and Tail Value at Risk values based on items and epcalc/eptype/eptype_tvar

write_ept_weighted(buffer, items, items_start_end, ...)

Generate Loss Exceedance Curve values and Tail Value at Risk values based on items and epcalc/eptype/eptype_tvar.

write_psept(buffer, items, items_start_end, ...)

Generate Per Sample Exceedance Probability Tables (PSEPT) for each individual sample, producing a separate loss

write_psept_weighted(buffer, items, items_start_end, ...)

Generate Per Sample Exceedance Probability Tables (PSEPT) for each individual sample, producing a separate loss

write_wheatsheaf_mean(buffer, mean_map, eptype, ...)

Generate Wheatsheaf Mean Exceedance Probability Table (EPT) by averaging losses for each return period

Module Contents

oasislmf.pytools.lec.aggreports.write_tables.get_loss(next_retperiod, last_retperiod, last_loss, curr_retperiod, curr_loss)[source]

Get loss based on current and next return period

Parameters:

  • next_retperiod (float) – Next return period

  • last_retperiod (float) – Previous return period

  • last_loss (float) – Previous Loss value

  • curr_retperiod (float) – Current return period

  • curr_loss (float) – Current Loss value

Returns:

Loss Value

Return type:

loss (float)

oasislmf.pytools.lec.aggreports.write_tables.fill_tvar(tail, tail_sizes, tail_offsets, summary_id, next_retperiod, tvar)[source]

Populate the Tail with retperiod and tvar values for summary_id

Parameters:

  • tail (ndarray[TAIL_valtype]) – Flat array of (return period, tvar) values

  • tail_sizes (ndarray[int64]) – Array of current fill size per summary_id

  • tail_offsets (ndarray[int64]) – Array of start positions per summary_id in tail

  • summary_id (int) – Summary ID

  • next_retperiod (float) – Next Return Period

  • tvar (float) – Tail Value at Risk

Returns:

Flat array of (return period, tvar) values tail_sizes (ndarray[int64]): Array of current fill size per summary_id

Return type:

tail (ndarray[TAIL_valtype])

oasislmf.pytools.lec.aggreports.write_tables.fill_tvar_wheatsheaf(tail, tail_sizes, tail_offsets, summary_id, sidx, num_sidxs, next_retperiod, tvar)[source]

Populate the Tail with retperiod and tvar values for (summary_id, sidx) pair

Parameters:

  • tail (ndarray[TAIL_valtype]) – Flat array of (return period, tvar) values

  • tail_sizes (ndarray[int64]) – Array of current fill size per (summary_id, sidx) idx

  • tail_offsets (ndarray[int64]) – Array of start positions per idx in tail

  • summary_id (int) – Summary ID

  • sidx (int) – Sample ID

  • num_sidxs (int) – Number of sidxs to consider

  • next_retperiod (float) – Next Return Period

  • tvar (float) – Tail Value at Risk

Returns:

Flat array of (return period, tvar) values tail_sizes (ndarray[int64]): Array of current fill size per (summary_id, sidx) idx

Return type:

tail (ndarray[TAIL_valtype])

oasislmf.pytools.lec.aggreports.write_tables.write_return_period_out(next_returnperiod_idx, last_computed_rp, last_computed_loss, curr_retperiod, curr_loss, summary_id, eptype, epcalc, max_retperiod, counter, tvar, tail, tail_sizes, tail_offsets, returnperiods, mean_map=None, is_wheatsheaf=False, num_sidxs=-1)[source]

Processes return periods and computes losses for a given summary, updating TVaR and mean map if required.

Parameters:

  • next_returnperiod_idx (int) – Index of the next return period to process.

  • last_computed_rp (float) – Last computed return period

  • last_computed_loss (float) – Last computed loss value

  • curr_retperiod (float) – Current return period being processed.

  • curr_loss (float) – Loss associated with the current return period.

  • summary_id (int) – Identifier for the current summary.

  • eptype (int) – Type of exceedance probability (0 = OEP, 1 = AEP).

  • epcalc (int) – Type of exceedance probability calculation.

  • max_retperiod (int) – Maximum return period to be used in calculations

  • counter (int) – Counter used for updating TVaR

  • tvar (float) – Tail Value at Risk

  • tail (ndarray[TAIL_valtype]) – Flat array of (return period, tvar) values

  • tail_sizes (ndarray[int64]) – Array of current fill size per summary/idx

  • tail_offsets (ndarray[int64]) – Array of start positions per summary/idx in tail

  • returnperiods (ndarray[np.int32]) – Return Periods array

  • mean_map (ndarray[MEANMAP_dtype], optional) – An array mapping used for mean loss calculations per Summary ID. Used for EPT output later. Defaults to None.

  • is_wheatsheaf (bool, optional) – If True, update the wheatsheaf TVaR structure.

  • num_sidxs (int, optional) – Number of sidxs to consider. Defaults to -1 if not is_wheatsheaf.

Returns:

Return period and Loss EPT data tail (ndarray[TAIL_valtype]): Flat array of (return period, tvar) values tail_sizes (ndarray[int64]): Array of current fill size per summary/idx last_computed_rp (float): Last computed return period last_computed_loss (float): Last computed loss value

Return type:

rets (list[EPT_dtype])

oasislmf.pytools.lec.aggreports.write_tables.write_tvar(epcalc, eptype_tvar, tail, tail_sizes, tail_offsets, max_summary_id)[source]

Get TVaR values for EPT output from tail

Parameters:

  • epcalc (int) – Type of exceedance probability calculation.

  • eptype_tvar (int) – Type of Tail Value-at-Risk (TVAR) to calculate (0 = OEP TVAR, 1 = AEP TVAR).

  • tail (ndarray[TAIL_valtype]) – Flat array of (return period, tvar) values

  • tail_sizes (ndarray[int64]) – Array of current fill size per summary_id

  • tail_offsets (ndarray[int64]) – Array of start positions per summary_id in tail

  • max_summary_id (int) – Maximum summary ID

Returns:

Return period and Loss EPT data

Return type:

rets (ndarray[EPT_dtype])

oasislmf.pytools.lec.aggreports.write_tables.write_tvar_wheatsheaf(num_sidxs, eptype_tvar, tail, tail_sizes, tail_offsets, total_idxs)[source]

Get TVaR values for PSEPT output from tail

Parameters:

  • num_sidxs (int) – Number of sidxs to consider.

  • eptype_tvar (int) – Type of Tail Value-at-Risk (TVAR) to calculate (0 = OEP TVAR, 1 = AEP TVAR).

  • tail (ndarray[TAIL_valtype]) – Flat array of (return period, tvar) values

  • tail_sizes (ndarray[int64]) – Array of current fill size per (summary_id, sidx) idx

  • tail_offsets (ndarray[int64]) – Array of start positions per idx in tail

  • total_idxs (int) – Total number of (summary_id, sidx) index entries

Returns:

Return period and Loss PSEPT data

Return type:

rets (ndarray[PSEPT_dtype])

oasislmf.pytools.lec.aggreports.write_tables.write_ept(buffer, items, items_start_end, max_retperiod, epcalc, eptype, eptype_tvar, use_return_period, returnperiods, max_summary_id, sample_size=1)[source]

Generate Loss Exceedance Curve values and Tail Value at Risk values based on items and epcalc/eptype/eptype_tvar

The loss calculation follows these principles: - For Aggregate Loss Exceedance Curves (AEP): The sum of all losses within a period is calculated. - For Occurrence Loss Exceedance Curves (OEP): The maximum loss within a period is taken. - TVAR (Tail Conditional Expectation): Calculated as the average of losses exceeding a given return period.

Parameters:

  • items (ndarray[LOSSVEC2MAP_dtype]) – Array mapping summary_id to loss value (and period_no/period_weighting where applicable)

  • items_start_end (ndarray[np.int32]) – An array marking where the start and end idxs are for each summary_id in the items array

  • max_retperiod (int) – Maximum return period to be used in calculations

  • epcalc (int) – Specifies the calculation method (mean damage loss, full uncertainty, per sample mean, sample mean).

  • eptype (int) – Type of exceedance probability (0 = OEP, 1 = AEP).

  • eptype_tvar (int) – Type of Tail Value-at-Risk (TVAR) to calculate (0 = OEP TVAR, 1 = AEP TVAR).

  • use_return_period (bool) – Use Return Period file.

  • returnperiods (ndarray[np.int32]) – Return Periods array

  • max_summary_id (int) – Maximum summary ID

  • sample_size (int, optional) – Sample Size. Defaults to 1.

  • buffer (ndarray[EPT_dtype]) – Pre-allocated output buffer that rows are written into and yielded from in chunks

Yields:

buffer (ndarray[EPT_dtype]) – Buffered chunks of EPT data

oasislmf.pytools.lec.aggreports.write_tables.write_ept_weighted(buffer, items, items_start_end, cum_weight_constant, epcalc, eptype, eptype_tvar, unused_period_weights, use_return_period, returnperiods, max_summary_id, sample_size=1)[source]

Generate Loss Exceedance Curve values and Tail Value at Risk values based on items and epcalc/eptype/eptype_tvar.

This function calculates weighted exceedance probability tables using cumulative period weightings (period_weighting), which impact the calculation of return periods. The weighting allows for more accurate representation of losses when event periods have different probabilities or frequencies of occurrence.

The loss calculation follows these principles: - For Aggregate Loss Exceedance Curves (AEP): The sum of all losses within a period is calculated. - For Occurrence Loss Exceedance Curves (OEP): The maximum loss within a period is taken. - TVAR (Tail Conditional Expectation): Calculated as the average of losses exceeding a given return period.

Parameters:

  • items (ndarray[LOSSVEC2MAP_dtype]) – Array mapping summary_id to loss value (and period_no/period_weighting where applicable)

  • items_start_end (ndarray[np.int32]) – An array marking where the start and end idxs are for each summary_id in the items array

  • cum_weight_constant (float) – Constant factor for scaling cumulative period weights.

  • epcalc (int) – Specifies the calculation method (mean damage loss, full uncertainty, per sample mean, sample mean).

  • eptype (int) – Type of exceedance probability (0 = OEP, 1 = AEP).

  • eptype_tvar (int) – Type of Tail Value-at-Risk (TVAR) to calculate (0 = OEP TVAR, 1 = AEP TVAR).

  • unused_period_weights (ndarray[float]) – Array of unused period weights

  • use_return_period (bool) – Use Return Period file.

  • returnperiods (ndarray[np.int32]) – Return Periods array

  • max_summary_id (int) – Maximum summary ID

  • sample_size (int, optional) – Sample Size. Defaults to 1.

  • buffer (ndarray[EPT_dtype]) – Pre-allocated output buffer that rows are written into and yielded from in chunks

Yields:

buffer (ndarray[EPT_dtype]) – Buffered chunks of EPT data

oasislmf.pytools.lec.aggreports.write_tables.write_psept(buffer, items, items_start_end, max_retperiod, eptype, eptype_tvar, use_return_period, returnperiods, max_summary_id, num_sidxs)[source]

Generate Per Sample Exceedance Probability Tables (PSEPT) for each individual sample, producing a separate loss exceedance curve for each sample, eptype, eptype_tvar.

Parameters:

  • items (ndarray[WHEATKEYITEMS_dtype]) – Array mapping (summary_id, sidx) to loss value (and period_no/period_weighting where applicable)

  • items_start_end (ndarray[np.int32]) – An array marking where the start and end idxs are for each (summary_id, sidx) pair in the items array

  • max_retperiod (int) – Maximum return period to be used in calculations

  • eptype (int) – Type of exceedance probability (0 = OEP, 1 = AEP).

  • eptype_tvar (int) – Type of Tail Value-at-Risk (TVAR) to calculate (0 = OEP TVAR, 1 = AEP TVAR).

  • use_return_period (bool) – Use Return Period file.

  • returnperiods (ndarray[np.int32]) – Return Periods array

  • max_summary_id (int) – Maximum summary ID

  • num_sidxs (int) – Number of sidxs to consider

  • buffer (ndarray[PSEPT_dtype]) – Pre-allocated output buffer that rows are written into and yielded from in chunks

Yields:

buffer (ndarray[PSEPT_dtype]) – Buffered chunks of PSEPT data

oasislmf.pytools.lec.aggreports.write_tables.write_psept_weighted(buffer, items, items_start_end, max_retperiod, eptype, eptype_tvar, unused_period_weights, use_return_period, returnperiods, max_summary_id, num_sidxs, sample_size, mean_map=None)[source]

Generate Per Sample Exceedance Probability Tables (PSEPT) for each individual sample, producing a separate loss exceedance curve for each sample, eptype, eptype_tvar.

Parameters:

  • items (ndarray[WHEATKEYITEMS_dtype]) – Array mapping (summary_id, sidx) to loss value (and period_no/period_weighting where applicable)

  • items_start_end (ndarray[np.int32]) – An array marking where the start and end idxs are for each (summary_id, sidx) pair in the items array

  • max_retperiod (int) – Maximum return period to be used in calculations

  • eptype (int) – Type of exceedance probability (0 = OEP, 1 = AEP).

  • eptype_tvar (int) – Type of Tail Value-at-Risk (TVAR) to calculate (0 = OEP TVAR, 1 = AEP TVAR).

  • unused_period_weights (ndarray[float]) – Array of unused period weights

  • use_return_period (bool) – Use Return Period file.

  • returnperiods (ndarray[np.int32]) – Return Periods array

  • max_summary_id (int) – Maximum summary ID

  • num_sidxs (int) – Number of sidxs to consider

  • sample_size (int) – Sample Size. Defaults to 1.

  • mean_map (ndarray[MEANMAP_dtype], optional) – An array mapping used for mean loss calculations per Summary ID. Used for EPT output later. Defaults to None.

  • buffer (ndarray[PSEPT_dtype]) – Pre-allocated output buffer that rows are written into and yielded from in chunks

Yields:

buffer (ndarray[PSEPT_dtype]) – Buffered chunks of PSEPT data

oasislmf.pytools.lec.aggreports.write_tables.write_wheatsheaf_mean(buffer, mean_map, eptype, epcalc, max_summary_id)[source]

Generate Wheatsheaf Mean Exceedance Probability Table (EPT) by averaging losses for each return period from a precomputed mean map.

Parameters:

  • mean_map (ndarray[MEANMAP_dtype]) – An array mapping used for mean loss calculations per Summary ID.

  • epcalc (int) – Specifies the calculation method (mean damage loss, full uncertainty, per sample mean, sample mean).

  • eptype (int) – Type of exceedance probability (0 = OEP, 1 = AEP).

  • max_summary_id (int) – Maximum summary ID

  • buffer (ndarray[EPT_dtype]) – Pre-allocated output buffer that rows are written into and yielded from in chunks

Yields:

buffer (ndarray[EPT_dtype]) – Buffered chunks of EPT data