--- a
+++ b/tests/conftest.py
@@ -0,0 +1,239 @@
+import logging
+import os
+import time
+from datetime import datetime
+
+import pandas as pd
+import pytest
+import spacy
+from helpers import make_nlp
+from pytest import fixture
+
+import edsnlp
+
+os.environ["EDSNLP_MAX_CPU_WORKERS"] = "2"
+os.environ["TZ"] = "Europe/Paris"
+
+try:
+    time.tzset()
+except AttributeError:
+    pass
+logging.basicConfig(level=logging.INFO)
+try:
+    import torch.nn
+except ImportError:
+    torch = None
+
+pytest.importorskip("rich")
+
+
+def pytest_collection_modifyitems(items):
+    """Run test_docs* at the end"""
+    first_tests = []
+    last_tests = []
+    for item in items:
+        if item.name.startswith("test_code_blocks"):
+            last_tests.append(item)
+        else:
+            first_tests.append(item)
+    items[:] = first_tests + last_tests
+
+
+@fixture(scope="session", params=["eds", "fr"])
+def lang(request):
+    return request.param
+
+
+@fixture(scope="session")
+def nlp(lang):
+    return make_nlp(lang)
+
+
+@fixture(scope="session")
+def nlp_eds():
+    return make_nlp("eds")
+
+
+@fixture
+def blank_nlp(lang):
+    if lang == "eds":
+        model = spacy.blank("eds")
+    else:
+        model = edsnlp.blank("fr")
+    model.add_pipe("eds.sentences")
+    return model
+
+
+def make_ml_pipeline():
+    nlp = edsnlp.blank("eds")
+    nlp.add_pipe("eds.sentences", name="sentences")
+    nlp.add_pipe(
+        "eds.transformer",
+        name="transformer",
+        config=dict(
+            model="prajjwal1/bert-tiny",
+            window=128,
+            stride=96,
+        ),
+    )
+    nlp.add_pipe(
+        "eds.ner_crf",
+        name="ner",
+        config=dict(
+            embedding=nlp.get_pipe("transformer"),
+            mode="independent",
+            target_span_getter=["ents", "ner-preds"],
+            span_setter="ents",
+        ),
+    )
+    ner = nlp.get_pipe("ner")
+    ner.update_labels(["PERSON", "GIFT"])
+    return nlp
+
+
+@fixture()
+def ml_nlp():
+    if torch is None:
+        pytest.skip("torch not installed", allow_module_level=False)
+    return make_ml_pipeline()
+
+
+@fixture(scope="session")
+def frozen_ml_nlp():
+    if torch is None:
+        pytest.skip("torch not installed", allow_module_level=False)
+    return make_ml_pipeline()
+
+
+@fixture()
+def text():
+    return (
+        "Le patient est admis pour des douleurs dans le bras droit, "
+        "mais n'a pas de problème de locomotion. "
+        "Historique d'AVC dans la famille. pourrait être un cas de rhume.\n"
+        "NBNbWbWbNbWbNBNbNbWbWbNBNbWbNbNbWbNBNbWbNbNBWbWbNbNbNBWbNbWbNbWBNb"
+        "NbWbNbNBNbWbWbNbWBNbNbWbNBNbWbWbNb\n"
+        "Pourrait être un cas de rhume.\n"
+        "Motif :\n"
+        "Douleurs dans le bras droit.\n"
+        "ANTÉCÉDENTS\n"
+        "Le patient est déjà venu pendant les vacances\n"
+        "d'été.\n"
+        "Pas d'anomalie détectée."
+    )
+
+
+@fixture
+def doc(nlp, text):
+    return nlp(text)
+
+
+@fixture
+def blank_doc(blank_nlp, text):
+    return blank_nlp(text)
+
+
+@fixture
+def df_notes():
+    N_LINES = 100
+    notes = pd.DataFrame(
+        data={
+            "note_id": list(range(N_LINES)),
+            "note_text": N_LINES * [text],
+            "note_datetime": N_LINES * [datetime.today()],
+        }
+    )
+
+    return notes
+
+
+def make_df_note(text, module):
+    from pyspark.sql import types as T
+    from pyspark.sql.session import SparkSession
+
+    data = [(i, i // 5, text, datetime(2021, 1, 1)) for i in range(20)]
+
+    if module == "pandas":
+        return pd.DataFrame(
+            data=data, columns=["note_id", "person_id", "note_text", "note_datetime"]
+        )
+
+    note_schema = T.StructType(
+        [
+            T.StructField("note_id", T.IntegerType()),
+            T.StructField("person_id", T.IntegerType()),
+            T.StructField("note_text", T.StringType()),
+            T.StructField(
+                "note_datetime",
+                T.TimestampType(),
+            ),
+        ]
+    )
+
+    spark = SparkSession.builder.getOrCreate()
+    notes = spark.createDataFrame(data=data, schema=note_schema)
+    if module == "pyspark":
+        return notes
+
+    if module == "koalas":
+        try:
+            import databricks.koalas  # noqa F401
+        except ImportError:
+            pytest.skip("Koalas not installed")
+        return notes.to_koalas()
+
+
+@fixture
+def df_notes_pandas(text):
+    return make_df_note(text, "pandas")
+
+
+@fixture
+def df_notes_pyspark(text):
+    return make_df_note(text, "pyspark")
+
+
+@fixture
+def df_notes_koalas(text):
+    return make_df_note(text, "koalas")
+
+
+@fixture
+def run_in_test_dir(request, monkeypatch):
+    monkeypatch.chdir(request.fspath.dirname)
+
+
+@pytest.fixture(autouse=True)
+def stop_spark():
+    yield
+    try:
+        from pyspark.sql import SparkSession
+    except ImportError:
+        return
+    try:
+        getActiveSession = SparkSession.getActiveSession
+    except AttributeError:
+
+        def getActiveSession():  # pragma: no cover
+            from pyspark import SparkContext
+
+            sc = SparkContext._active_spark_context
+            if sc is None:
+                return None
+            else:
+                assert sc._jvm is not None
+                if sc._jvm.SparkSession.getActiveSession().isDefined():
+                    SparkSession(sc, sc._jvm.SparkSession.getActiveSession().get())
+                    try:
+                        return SparkSession._activeSession
+                    except AttributeError:
+                        try:
+                            return SparkSession._instantiatedSession
+                        except AttributeError:
+                            return None
+                else:
+                    return None
+
+    session = getActiveSession()
+    if session is not None:
+        session.stop()