Feature/anthropic api#35

Merged

jpcompartir merged 57 commits intomainfrom

feature/anthropic_api

Dec 4, 2025

Commits on Dec 3, 2025

Commits on Dec 4, 2025

add endpointr_id to hf_build_request in prep for hf_embed_chunks
jpcompartir
committed
add the hf_embed_chunks function
jpcompartir
committed
update hf_embed_df to use the hf_embed_chunks function for efficiency
jpcompartir
committed
export hf_embed_chunks
jpcompartir
committed
fix merge conflicts in test-hf_embed.R
jpcompartir
committed
Fix part of out-of-date docs and examples
jpcompartir
committed
add the .handle_output_dir to start replacing .handle_output_file
jpcompartir
committed
bump version and add arrow to deps
jpcompartir
committed
add metadata.json to the output_dir - will be useful in the future for debugging people's code/errors (including my own)
jpcompartir
committed
add final alert for hf_embed_df function reporting number of successes/failures
jpcompartir
committed
start the hf_classify_chunks func
jpcompartir
committed
add input val to hf_classify_chunks
jpcompartir
committed
metadata and alerts etc. for hf_classify_chunks
jpcompartir
committed
fix typos in hf_classify_chunks
jpcompartir
committed
add texts to failures and successes for the hf_classify_chunks function
jpcompartir
committed
add the hf_classify_dev (with encrypted endpoints)
jpcompartir
committed
add the hf_get_model_max_length function for ammendments to hf_classify and later hf_embed
jpcompartir
committed
write inference parameters to metadata in hf_embed_df
jpcompartir
committed
add max_length parameter to hf_classify_df and write inference parameters to metadata
jpcompartir
committed
remove ... option for args passing in hf_classify_chunks/df
jpcompartir
committed
Remove max_length from hf_embed_df, and hf_classify_df - the solution is to turn on 'AUTO_TRUNCATE' in the Set up of the endpoint
jpcompartir
committed
add the 'hf_get_endpoint_info()` function to retrieve endpoint details
jpcompartir
committed
add @returns for hf_get_endpoint_info
jpcompartir
committed
Update test_embed tests following changes to file writing and arguments
jpcompartir
committed
Update README following changes to hf_*_df functions and move to chunks
jpcompartir
committed
add chunking and updated tests for hf_classify_*, similar to hf_embed_*
jpcompartir
committed
add new functions to _pkgdown.yml and new section for HF utilities
jpcompartir
committed
update the hugging_face_inference vignette in line with recent changes to the hf_*_df functions.
jpcompartir
committed
update roxygen2 docs for classify
jpcompartir
committed
Re-factor such that chunks doesn't overwrite variable names - writes to files with the origianl variable names.
jpcompartir
committed
rephrase to chunking vocab
jpcompartir
committed
refactor openai_embed.R for chunks
jpcompartir
committed
Move oai_embed_df to chunks func
jpcompartir
committed
fix tests affected by output_file -> output_dir changes
jpcompartir
committed
continue rebase
jpcompartir
committed
After refactoring clear that we actually want to return the .status as well as the .error and .error_msg, for easier filtering
jpcompartir
committed
Continue refactor with altering tests for .status expectation
jpcompartir
committed
Resolve any merge conflicts for .status change and refactors
jpcompartir
committed
update .create_error_tibble.Rd
jpcompartir
committed
Update news ahead of merge in feature branch and eventual version push
jpcompartir
committed
prevent httr2 from capturing error messages, as we've done in the refactor across oai_* funcs and hf_* funcs with the changes to basee_request (which ant doesn't use due to auth differences)
jpcompartir
committed
move funcs to .extract_api_error instead of manual plucking
jpcompartir
committed
Bring together some integration tests for when making big changes
jpcompartir
committed
switch to how we identify successes/failures - not using the httr2 funcs which check the httr2 request was successful/not
jpcompartir
committed