Add requirements for transform_prot (exclude on arm64 platforms as nm…

…slib install is broken)

setup.cfg

@@ -25,6 +25,7 @@ python_requires = >=3.7, <3.11
 install_requires =
     anaforatools~=1.1.0
     datasets~=2.4.0
+    en-core-sci-sm @ https://s3-us-west-2.amazonaws.com/ai2-s2-scispacy/releases/v0.5.1/en_core_sci_sm-0.5.1.tar.gz ; platform_machine!='arm64' and platform_machine!='aarch64'
     fastapi~=0.88.0
     filelock~=3.4.0
     huggingface-hub~=0.9.0
@@ -36,7 +37,9 @@ install_requires =
     PyRuSH~=1.0.3.6
     requests~=2.26.0
     scikit-learn~=1.0.2
+    scispacy==0.5.1; platform_machine!='arm64' and platform_machine!='aarch64'
     seqeval~=1.2.2
+    spacy~=3.4.0
     torch~=1.5
     transformers~=4.15
     uvicorn[standard]~=0.16.0

src/cnlpt/data/transform_prot.py

@@ -12,7 +12,14 @@
 
 TEST_DIR = "chemprot_development"
 TRAIN_DIR = "chemprot_training"
-nlp = spacy.load("en_core_sci_sm")
+
+nlp = None
+try:
+    import scispacy
+    nlp = spacy.load("en_core_sci_sm")
+except ImportError:
+    print(f"scispacy not installed; cannot run transform_prot.")
+    sys.exit(1)
 
 
 def remove_newline(review):
@@ -56,7 +63,10 @@ def get_intersect(ls1, ls2):
 
 
 def to_stanza_style_dict(text):
-    processed_doc = nlp(text)
+    if nlp is not None:
+        processed_doc = nlp(text)
+    else:
+        raise RuntimeError("nlp object not initialized but execution proceeded??")
 
     def sent_dict(spacy_sent):
         return [