port more test

Author: chengchingwen

Diff for: test/runtests.jl

@@ -3,7 +3,7 @@ using OhMyArtifacts
 using Dates
 using Test, Pkg
 
-using HuggingFaceApi: CONFIG_NAME, get_etag
+using HuggingFaceApi: CONFIG_NAME, get_etag, with_endpoint
 
 # https://github.com/huggingface/huggingface_hub/blob/f124f8be1e02ca9fbcda7a849e70271299ad5738/tests/testing_utils.py
 const DUMMY_MODEL_ID = "julien-c/dummy-unknown"
@@ -15,10 +15,24 @@ const DUMMY_MODEL_ID_PINNED_SHA256 =
 const SAMPLE_DATASET_IDENTIFIER = "lhoestq/custom_squad"
 const DATASET_ID = SAMPLE_DATASET_IDENTIFIER
 const DUMMY_DATASET_ID = "lhoestq/test"
+const DUMMY_DATASET_ID_REVISION_ONE_SPECIFIC_COMMIT =
+    "81d06f998585f8ee10e6e3a2ea47203dc75f2a16"
+
 const DATASET_REVISION_ID_ONE_SPECIFIC_COMMIT = "e25d55a1c4933f987c46cc75d8ffadd67f257c61"
 # One particular commit for DATASET_ID
 const DATASET_SAMPLE_PY_FILE = "custom_squad.py"
 
+# https://github.com/huggingface/huggingface_hub/blob/0c78398d42af1bb605b8d69c277b1751067d0d57/tests/testing_constants.py
+const USER = "__DUMMY_TRANSFORMERS_USER__"
+const FULL_NAME = "Dummy User"
+const PASS = "__DUMMY_TRANSFORMERS_PASS__"
+
+# Not critical, only usable on the sandboxed CI instance.
+const TOKEN = "hf_94wBhPGp6KrrTH3KDchhKpRxZwd6dmHWLL"
+
+const ENDPOINT_STAGING = "https://hub-ci.huggingface.co"
+
+
 @testset "HuggingFaceApi.jl" begin
     @test HuggingFaceURL(split("a/b/c",'/')...) == HuggingFaceURL("a", "b", "c")
     @test get_etag(HuggingFaceURL(DUMMY_MODEL_ID, CONFIG_NAME)) == DUMMY_MODEL_ID_PINNED_SHA1
@@ -59,6 +73,15 @@ const DATASET_SAMPLE_PY_FILE = "custom_squad.py"
 end
 
 @testset "Api endpoint" begin
+    info = with_endpoint(ENDPOINT_STAGING) do
+        HuggingFaceApi.whoami(TOKEN)
+    end
+    @test info.name == USER
+    @test info.fullname == FULL_NAME
+    @test info.orgs isa AbstractVector
+    valid_org_i = findfirst(org->org.name == "valid_org", info.orgs)
+    @test info.orgs[valid_org_i].apiToken isa AbstractString
+
     _api = HuggingFaceApi
     model_tags = _api.get_model_tags()
     for kind in ("library", "language", "license", "dataset", "pipeline_tag")
@@ -112,4 +135,102 @@ end
     @test length(d) == 1
     @test "huggingface" == d[1].author
     @test occursin("DataMeasurementsFiles", d[1].id)
+    d_raft =  _api.list_datasets(; filter = "benchmark:raft")
+    @test length(d_raft) > 0
+    @test "benchmark:raft" in d_raft[1].tags
+    d_lang_creat = _api.list_datasets(; filter = "language_creators:crowdsourced")
+    @test length(d_lang_creat) > 0
+    @test "language_creators:crowdsourced" in d_lang_creat[1].tags
+    d_lang_en = _api.list_datasets(; filter = "languages:en", limit=3)
+    @test length(d_lang_en) > 0
+    @test "languages:en" in d_lang_en[1].tags
+    d_lang_en_fr = _api.list_datasets(; filter = ["languages:en", "languages:fr"])
+    @test length(d_lang_en_fr) > 0
+    @test "languages:en" in d_lang_en_fr[1].tags
+    @test "languages:fr" in d_lang_en_fr[1].tags
+    d_multiling = _api.list_datasets(; filter = "multilinguality:unknown")
+    @test length(d_multiling) > 0
+    @test "multilinguality:unknown" in d_multiling[1].tags
+    d_sized = _api.list_datasets(; filter = "size_categories:100K<n<1M")
+    @test length(d_sized) > 0
+    @test "size_categories:100K<n<1M" in d_sized[1].tags
+    d_task = _api.list_datasets(; filter = "task_categories:audio-classification")
+    @test length(d_task) > 0
+    @test "task_categories:audio-classification" in d_task[1].tags
+    d_task_id = _api.list_datasets(; filter = "task_ids:automatic-speech-recognition")
+    @test length(d_task_id) > 0
+    @test "task_ids:automatic-speech-recognition" in d_task_id[1].tags
+    d_full = _api.list_datasets(; full = true)
+    @test length(d_full) > 100
+    @test any(d->haskey(d, :cardData), d_full)
+    d_author = _api.list_datasets(; author = "huggingface")
+    @test length(d_author) > 1
+    d_search = _api.list_datasets(; search = "wikipedia")
+    @test length(d_search) > 10
+    d_card = _api.list_datasets(; cardData = true)
+    @test count(d->haskey(d, :cardData), d_card) > 0
+    d_all = _api.list_datasets()
+    @test all(d->!haskey(d, :cardData), d_all) > 0
+
+    d_info = _api.dataset_info(DUMMY_DATASET_ID)
+    @test d_info.cardData isa AbstractDict && length(d_info.cardData) > 0
+    @test d_info.siblings isa AbstractVector && length(d_info.siblings) > 0
+    @test d_info.sha != DUMMY_DATASET_ID_REVISION_ONE_SPECIFIC_COMMIT
+    @test _api.dataset_info(DUMMY_DATASET_ID, revision=DUMMY_DATASET_ID_REVISION_ONE_SPECIFIC_COMMIT).sha ==
+        DUMMY_DATASET_ID_REVISION_ONE_SPECIFIC_COMMIT
+
+    mtr_all = _api.list_metrics()
+    @test length(mtr_all) > 10
+    @test any(m->haskey(m, :description), mtr_all)
+
+    m_author = _api.list_models(; author = "muellerzr")
+    @test length(m_author) > 0
+    @test occursin("muellerzr", m_author[1].modelId)
+    m_fb_bart = _api.list_models(; search = "facebook/bart-base")
+    @test occursin("facebook/bart-base", m_fb_bart[1].modelId)
+    m_fail = _api.list_models(; search = "muellerzr/testme")
+    @test length(m_fail) == 0
+    m_ms_tf = _api.list_models(; search = "microsoft/wavlm-base-sd", filter = "tensorflow")
+    @test length(m_ms_tf) == 0
+    m_ms_pt = _api.list_models(; search = "microsoft/wavlm-base-sd", filter = "pytorch")
+    @test length(m_ms_pt) > 0
+    m_task = _api.list_models(; search = "albert-base-v2", filter = "fill-mask")
+    @test "fill-mask" == m_task[1].pipeline_tag
+    @test occursin("albert-base-v2" , m_task[1].modelId)
+    @test length(_api.list_models(; filter = "dummytask")) == 0
+    @test length(_api.list_models(; filter = "en")) != length(_api.list_models(; filter = "fr"))
+    m_cplx = _api.list_models(; filter = ("text-classification", "pytorch", "tensorflow"))
+    @test length(m_cplx) > 1
+    @test all(m->"text-classification" == m.pipeline_tag || "text-classification" in m.tags, m_cplx)
+    @test all(m->"pytorch" in m.tags && "tensorflow" in m.tags, m_cplx)
+    @test all(m->haskey(m, :cardData), _api.list_models(filter="co2_eq_emissions", cardData = true))
+    @test all(m->!haskey(m, :cardData), _api.list_models(filter="co2_eq_emissions"))
+
+    s_all = _api.list_spaces(; full = true)
+    @test length(s_all) > 100
+    @test any(s->haskey(s, :cardData), s_all)
+    s_eval = _api.list_spaces(; author = "evaluate-metric")
+    @test ["evaluate-metric/trec_eval", "evaluate-metric/perplexity"] ⊆ [s.id for s in s_eval]
+    s_wiki = _api.list_spaces(; search = "wikipedia")
+    @test occursin("wikipedia", lowercase(s_wiki[1].id))
+    s_des = _api.list_spaces(; sort = "likes", direction = -1)
+    s_asc = _api.list_spaces(; sort = "likes")
+    @test s_des[1].likes > s_des[2].likes
+    @test s_asc[end-1].likes < s_asc[end].likes
+    @test length(_api.list_spaces(; limit=5)) == 5
+    s_bert = _api.list_spaces(; models = "bert-base-uncased")
+    @test "bert-base-uncased" in s_bert[1].models
+    s_d_wiki = _api.list_spaces(; datasets = "wikipedia")
+    @test "wikipedia" in s_d_wiki[1].datasets
+    s_link = _api.list_spaces(; linked = true)
+    @test any(s->haskey(s, :models), s_link)
+    @test any(s->haskey(s, :datasets), s_link)
+    @test any(s->haskey(s, :models) && haskey(s, :datasets), s_link)
+
+    @test length(with_endpoint(_api.list_datasets, ENDPOINT_STAGING)) <
+        length(with_endpoint(()->_api.list_datasets(TOKEN), ENDPOINT_STAGING))
+    @test length(with_endpoint(_api.list_models, ENDPOINT_STAGING)) <
+        length(with_endpoint(()->_api.list_models(TOKEN), ENDPOINT_STAGING))
+    @test length(with_endpoint(_api.list_spaces, ENDPOINT_STAGING)) <=
+        length(with_endpoint(()->_api.list_spaces(TOKEN), ENDPOINT_STAGING))
 end