Releases: ffverse/ffscrapr
tests-devel
Test files for devel branches
ffscrapr 1.4.8
An emergency patch to get this back onto CRAN, and packaging up some longstanding bugfixes also. Will need to revisit testing policy in future releases.
Fixes
- Fix
ff_league.mfl_conn()
bug which created duplicate rows for leagues with both an email auction and email draft. (resolves #349) - Fix hardcoded league year in all
ff_scoringhistory()
default arguments, now usesnflreadr::most_recent_season()
as the default upper bound.(resolves #347) - Require nflreadr v1.2.0+ due to data repository change.
dp_values()
now explicitly returns rather than invisibly returns.- Simplified GitHub repo to a GitHub Flow structure rather than a Git Flow structure for ease of maintenance.
- Re-export
nflreadr::ffverse_sitrep()
and encourage usage in bug reporting. - Fix
ff_transactions.mfl_conn()
bug that did not return auction bids, resolves #354 ff_league.espn_conn()
now counts TQB when evaluating qb_type and qb_count, resolves #367ff_starter_positions.mfl_conn()
now splits RB+WR and RB+WR+TE positions into their own positional rows and resummarisessleeper_connect()
now warns if you use a non-character league_id parameter - Sleeper league IDs are too long (18 characters) and longdouble support is inconsistent from OS to OSff_standings.mfl_conn()
changed to adapt to MFL API changes as documented in 2022 API release notes , resolves #366ff_starters.sleeper_conn()
bugfixed for API change (last-scored-leg -> leg)ff_scoring()
fixed for dev purrr issuesff_transactions.flea_conn()
bugfixed for infinite loop problem, resolves #356 (thank you @jdegregorio!)
ffscrapr 1.4.7
The main goal of v1.4.7 is to patch some minor bugs and to pre-emptively correct issues related to upstream dependency tidyr
.
Changes
- Fix
ff_scoringhistory.flea_conn()
kicking bug. - Fix
ff_schedule.mfl_conn()
NA bug. - Remove lubridate dep
- Update
.espn_stat_map()
to add threshold-based scoring and fix a missing defensive scoring setting - Update
ff_scoringhistory.espn_conn()
to add threshold calcs. (thank you, @scottfrechette!) - Fix
espn_potentialpoints()
to move regular flex ahead of superflex ff_scoring()
for MFL gets some tweaks to support changes to tidyr (thanks, @DavisVaughan!)
Thanks
Grateful to @davisderodes,@DavisVaughan, @draftydesert277, @scottfrechette, and @TheMathNinja for their contributions to this release - feedback, issues, and PRs!
ffscrapr 1.4.6
The main goal of v1.4.6 is to patch some minor bugs, introduce a unified dependency on nflreadr, and add some support for kicker scoring.
New features
ff_template()
adds a template connection forff_starter_positions()
,ff_scoring()
, andff_scoringhistory()
to more easily pull common scoring templates.- Add kicker stat mappings for
ff_scoringhistory()
and support ~ most kicker scorings (MFL's may be weird.)
Minor changes
ff_league()
for MFL now returns league type ("dynasty", "keeper", "redraft") if the field is set.ff_league()
for MFL now returns draft type (e.g. "email draft", "live auction") if the field is set.ff_league()
for MFL now returns draft player pool (e.g. "Rookie", "Veteran", "Both") if the field is set. (These three changes fix #311, thanks @maat7043!)ff_schedule()
for ESPN now credits a bye team with a win instead of causing an error.espn_getendpoint()
no longer returns an error for seasons prior to 2018. (Thanks, @JoeSydlowski!)nflfastr_weekly()
andnflfastr_roster()
now use the nflreadr package functions back-end- Added some names to the name database, c/o @SCasanova via #320 - thank you, Santiago!
ff_starters()
for ESPN now also tries to return a projected score for the week - thank you, @scottfrechette! Resolves #323.ff_schedule()
for Sleeper no longer broken - resolves #322.ff_schedule()
for MFL now returns NA as the result for unplayed games and becomes consistent with other platforms.ff_scoringhistory()
for MFL now parses decimal ranges correctly- Refactor
ff_scoringhistory()
into memoisable pieces (.nflfastr_offense_long
,.nflfastr_kicker_long
,.nflfastr_roster
) - Update http error messages for user readability.
- Minimum R version bumped to 3.6.0 - nflreadr uses RDS versions that require a minimum of R3.6.
Thanks
Immensely grateful to: Joe Sydlowski, Matt Spencer, Santiago Casanova, Scott Frechette, and all others who raised issues and contributed code towards this release!
ffscrapr 1.4.5
The main goal of v1.4.5 is to comply with CRAN somehow testing our \donttest
examples.
Minor changes
ff_league()
for ESPN now works for non-ppr leagues.- Include
try()
in all examples because apparently\donttest
doesn't do anything anymore 😡 (fixes #308)
ffscrapr 1.4.4
ffscrapr v1.4.4 patches some bugs, extends a new helper for removing html tags, extends features to support a new package (ffsimulator) and cuts over to a new GitHub org and website domain.
Minor changes
mfl_getendpoint()
and similar get_endpoint functions have an improved print method that tells you whether the request was successful.- Added errorhandling for ESPN methods of
ff_starters()
,ff_transactions()
,ff_rosters()
,espn_potentialpoints()
,ff_draft()
. Fixes #297, thanks for the report @jpiburn! - Added message to warn if memoise is turned off, onAttach.
ff_transactions()
now correctly handles leagues using waiver priority, fixes #299 - thanks for the report @BarkovMVP!ff_league()
for Sleeper now identifies best ball leagues, fixes #300ff_scoring()
for MFL now has apoints_type
column that is either "each" or "once" - this helps delineate fantasy points awarded for meeting thresholds/bonuses from points awarded per-stat, fixes #301ff_scoringhistory()
for MFL applies the newff_scoring()
points_type
column to calculate scoring history. This should result in more sensible results for leagues with bonus scoring! Fixes #301.- Add more mismatch names (Michael -> Mike Vick, Chris Wells -> Beanie Wells)
- Add
dp_clean_html()
to clean html from names (cough MFL cough) ff_starter_positions()
now handles a "range" inside of the total_starters column for MFL, resolves #304- Cutover to ffverse.com domain and github organization
ffscrapr 1.4.3
The main goal of v1.4.3 is to patch some minor bugs.
Minor changes
ff_league()
now has a column that returns the platform's season - this is most useful when running ff_league in batch. Thank you to @joeflan for the contribution! (Fixes #287)- Added helper script in dev/ for switching between test cache versions.
- Added helper script in dev/ for rebuilding test cache.
ff_schedule()
for Sleeper now should extend into playoff weeks. (Fixes #289)ff_draft()
for Sleeper now has an auction_amount column if it is an auction draft. (Fixes #291)dp_cleannames()
now can flip names originally presented in LastName, FirstName order into First Last, and also has a custom name database to convert common alternate names to a unified one.nflfastr_stat_mapping
anddp_name_mapping
are namespaced within the package, so that they can be used internally and externally a little more robustly.- Rebuilt test cache in ffscrapr-tests.
ffscrapr 1.4.2
The main goal of v1.4.2 is to patch some minor bugs.
Minor changes
espn_rosters
now has aweek
argument to return a specific week of data. (Fixes #276, thanks @scottfrechette!)ff_scoringhistory
now tries to return a platform-specific player ID, where available. (Fixes #275)- Roxygen configured to use markdown (with the roxygen2md package) - hopefully fixes the lifecycle badge issues?
sleeper_players()
- gsis_id getsstr_squish
applied to it to resolve some stowaway whitespaces.mfl_players()
- addany_of
handling for columns to handle column name changes.sleeper_transactions
now handles offseason transactions correctly
ffscrapr 1.4.1
The main goal of v1.4.1 is to patch some issues raised by CRAN checks and also correct some bugs in the new experimental ff_scoringhistory
and ff_starter_positions
functions released in v1.4.0.
New features
nflfastr_stat_mapping
is a dataframe that maps nflfastr columns to fantasy scoring rules, and is now exported for end-user usage. It is primarily used inside offf_scoringhistory()
.
Minor changes
- Added
release_questions
to help remind me to maintain test infrastructure mfl_starter_positions
now correctly calculates offensive starters (first via "iop_starters" if defined and then otherwise by subtracting idp starters) as well as kdst_starters (Resolves #264)- Redirected nflfastr download functions to the new nflverse/ repository locations. (Resolves #268)
.ffscrapr_env
relocated from being a child of the base environment to being a child of the empty environment (Resolves #269)ff_scoringhistory
refactored and adds better support for MFL's fumbles and firstdowns. (Resolves #265)
ffscrapr 1.4.0
The main goal of v1.4.0 is to add preliminary support for connecting ffscrapr to nflfastR weekly data, and to clean up bugs from v1.3.0. Huge thanks goes to Joe Sydlowski for his contributions on scoring history (and everything else DynastyProcess!)
New Features
nflfastr_weekly()
imports weekly offensive statistics from nflfastR's data repository.nflfastr_rosters()
imports team rosters from nflfastR's roster repository.ff_scoringhistory()
connects your league's scoring settings to the nflfastr data (c/o the functions above), and allows you to reconstruct historical scoring for your league!ff_starter_positions()
describes the starter rules for each player/position, including min and max starters at each position accounting for flex spots. This should be useful for calculating things like value over replacement!
Breaking Changes
ff_scoring
for ESPN loses the "override_pos" which only had a value of 16 for D/ST specific rulesff_scoring
rules for ESPN and Sleeper have been expanded from one line per rule to one line per position per rule
Minor Changes
- Force
mfl_playerscores
to use season + league specific players call where possible (#239) - Reduce minimum rows for
flea_rosters
test to 200, which fixes an API check issue (#242) - Switching memoise backend to cachem package, fixes CRAN check issue where digest is no longer imported by memoise. Cachem also apparently more performant! (#244)
sleeper_transactions
now correctly handles multiple dropped players in one transaction. (#246).mfl_transactions
- adds numeric string parsing to fix bug in auction bid amount (#)mfl_draft
now calculates age as of timestamp and also adds an overall column. (#259, #260)