Start Yahoo Implementation #336 #424

caltonji · 2023-12-21T20:13:30Z

Starting work on: #336

This PR implements ff_connect, ff_userleagues, and ff_franchises for Yahoo.

To login, users will need to

Obtain your league_id from any yahoo URL in your leagues. In this link it's 77275: https://football.fantasysports.yahoo.com/f1/77275/10
Obtain an auth token by signing into Yahoo here https://lemon-dune-0cd4b231e.azurestaticapps.net/.

I built that website for grabbing the auth token as part of a past project https://introductory.medium.com/download-yahoo-fantasy-football-data-bc1e1db7ee49

I plan to finish the Yahoo implementation, but I'm starting small.

caltonji · 2023-12-21T20:14:51Z

R/yahoo_franchises.R

+  franchise_name <- xml2::xml_text(xml2::xml_find_all(xml_doc, "//team/name"))
+  user_name <- xml2::xml_text(xml2::xml_find_all(xml_doc, "//manager/nickname"))
+  user_id <- xml2::xml_text(xml2::xml_find_all(xml_doc, "//manager/guid"))
+  co_owners <- NA # This is available in Yahoo but I'm bad at R so I can't tell what format it's supposed to be.


What format should the co_owners field be in?

Hmm. Generally have provided just the co-manager IDs in the past as a nested vector, can do that here or provide more of these details collapsed into a nested list/tibble.

Studying the example provided here and I'm not sure your above code will work as intended because xml_find_all will strip every user (manager or co-manager) into a plain vector.

This might result in inaccurate output when combined into a dataframe later (i.e. if franchise_name is a vector of 12 names and user_name is a vector of 16 names, combining these back into a base R dataframe would result in recycling the shorter vector to match the longer one)

Can assist further if you provide the full XML output from xml_doc?

You're right. Hadn't tested with co_owners yet. I updated it to support co_owners as a nested list. I also did add an example xml_file tests/yahoo_example_responses/single-league-teams.xml

tanho63

Happy to see someone take this on! I've added a feature branch yahoo which we can use to merge reviewed progress to (so that any other platform PRs don't affect this one).

Have left a few comments, feel free to reply here or take it to discord DMs for discussions if preferred

R/yahoo_api.R

R/yahoo_connect.R

R/yahoo_api.R

R/yahoo_connect.R

R/yahoo_franchises.R

tanho63 · 2023-12-22T01:53:39Z

R/yahoo_franchises.R

+  franchise_name <- xml2::xml_text(xml2::xml_find_all(xml_doc, "//team/name"))
+  user_name <- xml2::xml_text(xml2::xml_find_all(xml_doc, "//manager/nickname"))
+  user_id <- xml2::xml_text(xml2::xml_find_all(xml_doc, "//manager/guid"))
+  co_owners <- NA # This is available in Yahoo but I'm bad at R so I can't tell what format it's supposed to be.


Hmm. Generally have provided just the co-manager IDs in the past as a nested vector, can do that here or provide more of these details collapsed into a nested list/tibble.

tanho63 · 2023-12-22T01:54:39Z

R/yahoo_franchises.R

+  franchise_name <- xml2::xml_text(xml2::xml_find_all(xml_doc, "//team/name"))
+  user_name <- xml2::xml_text(xml2::xml_find_all(xml_doc, "//manager/nickname"))
+  user_id <- xml2::xml_text(xml2::xml_find_all(xml_doc, "//manager/guid"))
+  co_owners <- NA # This is available in Yahoo but I'm bad at R so I can't tell what format it's supposed to be.


Studying the example provided here and I'm not sure your above code will work as intended because xml_find_all will strip every user (manager or co-manager) into a plain vector.

This might result in inaccurate output when combined into a dataframe later (i.e. if franchise_name is a vector of 12 names and user_name is a vector of 16 names, combining these back into a base R dataframe would result in recycling the shorter vector to match the longer one)

Can assist further if you provide the full XML output from xml_doc?

R/yahoo_userleagues.R

caltonji

Thank you for the review @tanho63! I tried to address all feedback.

R/yahoo_api.R

R/yahoo_connect.R

caltonji · 2023-12-28T07:36:18Z

R/yahoo_franchises.R

+  franchise_name <- xml2::xml_text(xml2::xml_find_all(xml_doc, "//team/name"))
+  user_name <- xml2::xml_text(xml2::xml_find_all(xml_doc, "//manager/nickname"))
+  user_id <- xml2::xml_text(xml2::xml_find_all(xml_doc, "//manager/guid"))
+  co_owners <- NA # This is available in Yahoo but I'm bad at R so I can't tell what format it's supposed to be.


You're right. Hadn't tested with co_owners yet. I updated it to support co_owners as a nested list. I also did add an example xml_file tests/yahoo_example_responses/single-league-teams.xml

R/yahoo_userleagues.R

R/yahoo_franchises.R

tanho63

Nice work! I added some test mocks to your PR branch so that we can start using some tests here and added some crude tests to tests/testthat/tests-yahoo.R. Might be easiest to do up the PR, add the tests, and then send me a token to do the mock-test-uploads 😀

Left a few style notes for a more "tidy" way to do this, mostly-optional so if you'd prefer I merge and we tackle that after we get a more complete first implementation let me know

R/yahoo_connect.R

tanho63 · 2024-01-12T03:58:52Z

R/yahoo_franchises.R

+ff_franchises.yahoo_conn <- function(conn) {
+  glue::glue("leagues;league_keys={conn$league_key}/teams") %>%
+    yahoo_getendpoint(conn) %>%
+    {


ditto with above re: getElement etc etc

tanho63 · 2024-01-12T04:13:54Z

R/yahoo_franchises.R

+    }
+}
+
+.yahoo_process_franchises_response <- function(xml_doc) {


Token you gave me expired before I could fully flesh out this example, but here's an alternative way to approach this problem - convert xml to list and use tidyr/dplyr/purrr to unnest out the dataframe you want

x <- xml_doc %>% xml2::as_list() %>% purrr::pluck("fantasy_content", "leagues","league","teams") %>% tibble::tibble() %>% tidyr::unnest_wider(1) %>% dplyr::select( franchise_id = team_id, franchise_name = name, division_id = division_id, managers ) %>% tidyr::unnest_longer(c(franchise_id, franchise_name, division_id)) %>% dplyr::mutate( managers = purrr::map(managers, ~ tibble::tibble(.x) %>% tidyr::unnest_wider(1)) ) %>% tidyr::hoist( managers, "user_name" = "nickname", "user_id" = "guid", "user_felo" = "felo_score" ) %>% tidyr::unnest_longer(c(user_name, user_id, user_felo)) %>% tidyr::unnest_longer(c(user_name, user_id, user_felo))

I spent a few hours reading up on the tidyverse but I just can't for the life of me figure out how to work with the as_list xml output. I'm going to go with the xml_find_all solutions for now since they're super intuitive for the uninitated.

the uninitated.

the un-tidy-doctrinated 😂

That's fine. Let's put in getElement in all the relevant places and call it the finish line for this PR?

Maybe useful future reading if you want to look at unnesting in the future: https://r4ds.hadley.nz/rectangling

Co-authored-by: Tan Ho <38083823+tanho63@users.noreply.github.com>

caltonji commented Dec 21, 2023

View reviewed changes

tanho63 changed the base branch from main to yahoo December 22, 2023 00:35

tanho63 reviewed Dec 22, 2023

View reviewed changes

caltonji force-pushed the yahoo-first-pr branch from 5f22bf8 to 19bf511 Compare December 24, 2023 15:08

caltonji commented Dec 28, 2023

View reviewed changes

filter out empty drafts from sleeper ff_draft (ffverse#425)

4b61a17

caltonji requested a review from tanho63 January 11, 2024 16:41

tanho63 approved these changes Jan 12, 2024

View reviewed changes

tanho63 and others added 15 commits January 18, 2024 11:34

Update R-CMD-check.yaml

42a7e84

Add yahoo_connect and yahoo_franchises

e0aac3b

Add autogenerated documentation

3086b0c

Add support for ff_connect

7569993

Run styler

4892324

Added support for co_owners

779ca8a

Update styling. Switch userleagues to be more R like

77631c3

Switch to asking for season

3c2ada0

Switch to API call for game_id

c733fec

Small fixes

8a6f143

add some mocked tests

98f5e78

remove bypass of local_mock_api()

3a73277

Update R/yahoo_connect.R

8289075

Co-authored-by: Tan Ho <38083823+tanho63@users.noreply.github.com>

use getElement

baff51d

getElement

07e1d9d

caltonji force-pushed the yahoo-first-pr branch from e96f36d to 07e1d9d Compare January 19, 2024 05:53

tanho63 merged commit 4b9c680 into ffverse:yahoo Jan 28, 2024
0 of 5 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Start Yahoo Implementation #336 #424

Start Yahoo Implementation #336 #424

caltonji commented Dec 21, 2023 •

edited

Loading

caltonji Dec 21, 2023

tanho63 Dec 22, 2023

tanho63 Dec 22, 2023

caltonji Dec 28, 2023

tanho63 left a comment

tanho63 Dec 22, 2023

tanho63 Dec 22, 2023

caltonji left a comment

caltonji Dec 28, 2023

tanho63 left a comment

tanho63 Jan 12, 2024

tanho63 Jan 12, 2024

caltonji Jan 17, 2024

tanho63 Jan 18, 2024 •

edited

Loading

tanho63 Jan 18, 2024 •

edited

Loading

Start Yahoo Implementation #336 #424

Start Yahoo Implementation #336 #424

Conversation

caltonji commented Dec 21, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tanho63 left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

caltonji left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tanho63 left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tanho63 Jan 18, 2024 • edited Loading

Choose a reason for hiding this comment

tanho63 Jan 18, 2024 • edited Loading

Choose a reason for hiding this comment

caltonji commented Dec 21, 2023 •

edited

Loading

tanho63 Jan 18, 2024 •

edited

Loading

tanho63 Jan 18, 2024 •

edited

Loading