fee overhaul & targeted long fix

[dpaiton]

Description

This PR is the result of an audit of the fee functions. The following changes were made (one commit per bullet):

1. revert the rounding adjustment from Fix open_short_curve_fee #97. We realized that the bug actually lies in the solidity test utils. This has been fixed so I reverted the rounding behavior to match.
2. update docstrings to use a more consistent format. I introduce a convention of my own to use upper-case $\Phi_{c,ol}(...)$ for the fee function and lower-case $\phi_{c}$ for the fee multiplier constant. The subscript on the fee function is the fee type $\in c, g, f$ for curve, gov, and flat; as well as the trade type $\in ol, cl, os, cs$ for opening or closing a long or short. This replaces the inconsistent method of using e.g. c(x) at some times and curve_fee at other times, or using repeat symbols for different equations.
3. add maybe_curve_fee arg to gov fee calculations to fix Avoid redundant calculations for calculate_close_short_flat_plus_curve #43 and reduce computation.
4. change the rounding behavior to match solidity (note it's matching the eqs in internal/HyperdriveBase.sol
5. the above changes revealed a unit issue in calculate_targeted_long where the gov fee (and derivative) was not converted to shares. I updated my overleaf doc, the docstrings, and the code itself.

Review Checklists

Please check each item before approving the pull request. While going
through the checklist, it is recommended to leave comments on items that are
referenced in the checklist to make sure that they are reviewed.

• Testing
  • Are there new or updated unit or integration tests?
  • Do the tests cover the happy paths?
  • Do the tests cover the unhappy paths?
  • Are there an adequate number of fuzz tests to ensure that we are
    covering the full input space?
  • If matching Solidity behavior, are there differential fuzz tests that
    ensure that Rust matches Solidity?

[jrhea]

lgtm