Directdm

Fixed
- Key error in wilson_coefficients.py (`cNR13n` was not defined)
- Removed spurious NLO terms in `cNR` coefficients
- Deleted spurious dimension-seven Wilson coefficients. The basis now corresponds to [arxix [v3] of this paper](https://arxiv.org/abs/1710.10218).

Changed
- `example.py` includes a few more examples for RG running below the weak scale

Fixed
- Inconsistent implementation of double weak insertions for lepton operators
- Typos in matching of dimension-seven Wilson coefficients (matching to wrong NR coefficient, typos in form factors)
- A bug in NLO option: cNR coefficients were partially overwritten
- decoupling alphas from 4 to 3 flavors at 2 GeV instead of mc(mc)
- Error in QED anomalous dimension for tensor / scalar operators

Changed
- Updated charm-quark mass to most recent PDG value
- Removed internal options "DOUBLE_WEAK" and "double_QCD"
- class `WC_3f`: Removed method `run` and corresponding option `RGE`
- Globally removed the option to set the lower running scale
- Moved definitions of more dependent parameters to num_input

Added
- Additional dependent parameters (masses and strong coupling at various scales) to num_input
- NR coefficients `cNR13p`, `cNR13n`, `cNR14p`, `cNR14n`
- "slope of strange vector current" contributions for double weak insertions

Fixed
- Typos in expressions for cNR (no impact on numerics)

Changed
- Updated USAGE.md to include all input parameters
- Removed default input dictionary from single nucleon form factor classes

Added
- The strange electric charge radius to the input parameters
- The slope of the vector current form factor at zero momentum transfer (only strange quark)
- Missing NLO terms for the strange vector current

Added
- The chiral NLO terms for the strange-quark vector current operators (these are leading since the LO terms vanish)

Fixed
- Inconsistencies in implementation of dependent input parameters

Changed
- Updated numerical values of input parameters according to PDG 2020 and most recent lattice results
- Only "primary" input parameters can be changed by the user (see USAGE.md for more details). All dependent parameters are calculated automatically.
An invalid key in the input dictionary will cause the code to abort. This is maybe somewhat drastic, but
otherwise the code would run without the updated parameter which is probably not intended either.
- The QCD and electroweak MSbar top-quark mass at mu = MZ is now consistently used in the numerics.
The approximate relations for converting the renormalization scheme are taken from [this reference](https://arxiv.org/abs/1212.4319).

Added
- Some four-loop relations for running and decoupling; just for completeness and convenience;
they are currently not used in the code

Fixed
- Normalization of nonrelativistic coefficients for scalar DM (thanks to Marco Fedele for pointing this out!)