v0.7.1

Fixes:

• pyscallist now allows dataframes with index not starting at zero.
• Logging is not configured in each module, should only be set by client scripts

Changes:

• Drop Python 2 support

v0.7.0

Added features:

• Monotonocity is always enforced (in the last digit) in relative permeability data (previously only enforced for Pc)
• swl can now be initialized from height-above-free-water-level, if simple J parameters are included in input table.
• swcr can be initialized relative to swl, by using the swcr_add parameter in the input table.

Fixes:

• Interpolations between relperm curves originating in tables and not from parameters have improved endpoint detection.

v0.6.1

Bugs fixed:

• Resolves crash due to floating point representation error when interpolating from dataframes for gasoil.

New features:

• --version option to pyscal command line client.

v0.6.0

New features:

• GasWater is supported.

Changed behaviour:

• In Pyscal < 0.6.0 and WaterOil, the oil relperm curve shape started with kro(swcr) = kroend. Now it starts with kro(swl) = kroend. Thus, if swcr > swl, your oil relperm will change when moving to Pyscal 0.6.x.
• In Pyscal < 0.6.0 and GasOil, the oil relperm curve shape started with kro(sgcr) = kroend. Now it starts with kro(0) = kroend. Thus, if sgcr > 0 (more common than swcr > swl), your oil relperm will change when moving to Pyscal 0.6.x. The oil relative permeability will in general be lowered through this change, and the change can be significant for large sgcr.
• kromax is not any longer needed, ignored and deprecated
• krogend and krowend are deprecated parameter names, merged into the parameter name kroend. krogend/krowend can still be used but warnings will be issued.

Bugfixes:

• The string "pessimistic" was not recognized due to typo, in input xlsx sheets.

v0.5.1

Bugfix for command line client on oil-water problems, resolving crash.

v0.5.0

Add saturation point between sorw and 1 in WaterOil saturation tables.

These saturation points were previously ignored (i.e. not explicitly include), as the correlations imply linear relative permeability curves in this section, and Eclipse will interpolate linearly anyway. But the saturation points may be important to resolve initialization from capillary pressure, and the capillary pressure should not be interpolated linearly in this region.

Most SWOF tables will now have more rows, and could trigger necessary changes in NSSFUN in TABDIMS in cases where the table length settings are already set minimal. The default in E100 is 20, which is already too low for standard pyscal output.

v0.4.3

Bugfix release:

• Interpolated relperms values was seen (in corner-case) to be non-monotone due to IEEE754 representation errors, and due to default float-printing being truncation. Fixed by an extra rounding step before printing to ASCII tables.
• Capillary pressure was seen in a different corner-case to be not strict-monotone in the ASCII representation with 7 digits. Fix by brute forcing non-zero pc vectors to be stricly monotone for the given printed numerical accuracy.
• Gasoil capillary pressure is always increasing, but gasoil.from_table() would refuse to accept anything but decreasing, fixed.

v0.4.2

Minor release:

• Robustified parsing relperm data from file.

v0.4.1

Minor feature release:

• Add petrophysical version of simple J-function for capillary pressure
• Add figures in documentation explaining saturation and relperm endpoints
• Allow aliases for low and high in the CASE column for SCAL recommendations.
• Added getitem for PyscalList objects.
• Preserve tags through interpolations and SCAL recommendations
• Ability to dump a merged DataFrame of all tables in a PyscalList

v0.4.0

Feature release:

• Added command line tool pyscal accepting CSV or XLSX as input with relative permeability parameters.

Bugs fixed:

• Cosmetic in pc-string-generation
• Some logger warnings were wrong.
• Avoid numerical corner case where sgn collapses.
• Don't give so many warnings for unused values when they are defaulted.