Limit number of requests sent to OpenAlex in tests (asreview#44)

Author: PeterLombaers

tests/demo_data/snowballing_doi.csv

@@ -1,3 +1,3 @@
 ,title,doi,included
-0,"Social Networks Analysis: Tools, Measures and Visualization",https://doi.org/10.1007/978-1-4471-4054-2_1,1
-1,"Genome-wide Association Study of Alcohol Dependence",https://doi.org/10.1001/archgenpsychiatry.2009.83,0
+0,"Myrmecochorous plants in Australia and their dispersal by ants",https://doi.org/10.1071/bt9750475,1
+1,"Mimicking the one-dimensional marginal distributions of processes having an ito differential",https://doi.org/10.1007/bf00699039,0

tests/demo_data/snowballing_openalex.csv

@@ -1,3 +1,3 @@
 ,openalex_id,title,included
-0,https://openalex.org/W2234238252,"Social Networks Analysis: Tools, Measures and Visualization",1
-1,https://openalex.org/W1977467968,"Genome-wide Association Study of Alcohol Dependence",0
+0,https://openalex.org/W2051970045,"Myrmecochorous plants in Australia and their dispersal by ants",1
+1,https://openalex.org/W104454400,"Mimicking the one-dimensional marginal distributions of processes having an ito differential",0

tests/test_snowball.py

@@ -1,13 +1,45 @@
 from pathlib import Path
 
 import pandas as pd
+import pyalex
 
 from asreviewcontrib.datatools.snowball import backward_snowballing
 from asreviewcontrib.datatools.snowball import forward_snowballing
 from asreviewcontrib.datatools.snowball import openalex_from_doi
 from asreviewcontrib.datatools.snowball import snowball
 
 INPUT_DIR = Path(__file__).parent / "demo_data"
+EMAIL = "asreview@uu.nl"
+
+pyalex.config.email = EMAIL
+
+# These works were chosen for testing forward snowballing.
+# They have a DOI, they cite and are cited by, their cited_by_count is less than 400,
+# so it takes only two requests to get all citing works. And they are from the previous
+# century so the cited_by_count is unlikely to change very much.
+# These are also the same records as in the demo datasets 'snowballing_doi.csv' and
+# 'snowballing_openalex.csv'.
+WORKS = [
+    {
+        "id": "https://openalex.org/W2051970045",
+        "doi": "https://doi.org/10.1071/bt9750475",
+        "title": "Myrmecochorous plants in Australia and their dispersal by ants",
+        "cited_by_count": 372,
+        "cited_by": "https://openalex.org/W2174650845",
+        "cites": "https://openalex.org/W1538725992",
+    },
+    {
+        "id": "https://openalex.org/W104454400",
+        "doi": "https://doi.org/10.1007/bf00699039",
+        "title": (
+            "Mimicking the one-dimensional marginal distributions of processes having"
+            " an ito differential"
+        ),
+        "cited_by_count": 299,
+        "cited_by": "https://openalex.org/W1842249978",
+        "cites": "https://openalex.org/W1513091520",
+    },
+]
 
 
 def test_openalex_from_doi():
@@ -41,32 +73,30 @@ def test_backward_snowballing():
 
 
 def test_forward_snowballing():
-    identifiers = [
-        "https://openalex.org/W4281483266",
-        "https://openalex.org/W2008620264",
-    ]
+    identifiers = [work["id"] for work in WORKS]
 
     forwards_citations = forward_snowballing(identifiers)
 
-    assert "https://openalex.org/W4386305682" in [
+    assert WORKS[0]["cited_by"] in [
         field_dict["id"] for field_dict in forwards_citations[identifiers[0]]
     ]
-    assert "https://openalex.org/W2124637492" in [
+    assert WORKS[1]["cited_by"] in [
         field_dict["id"] for field_dict in forwards_citations[identifiers[1]]
     ]
 
 
 def test_openalex_id_forward(tmpdir):
-    out_fp = Path(tmpdir, "forward_all.csv")
+    out_fp = Path(tmpdir, "forward.csv")
     snowball(
         input_path=INPUT_DIR / "snowballing_openalex.csv",
         output_path=out_fp,
         forward=True,
         backward=False,
         use_all=False,
+        email=EMAIL,
     )
     df = pd.read_csv(out_fp)
-    assert len(df) >= 23
+    assert len(df) >= 364
 
     all_out_fp = Path(tmpdir, "forward_all.csv")
     snowball(
@@ -75,22 +105,24 @@ def test_openalex_id_forward(tmpdir):
         forward=True,
         backward=False,
         use_all=True,
+        email=EMAIL,
     )
     df_all = pd.read_csv(all_out_fp)
-    assert len(df_all) >= 387
+    assert len(df_all) >= 656
 
 
 def test_openalex_id_backward(tmpdir):
-    out_fp = Path(tmpdir, "forward_all.csv")
+    out_fp = Path(tmpdir, "backward.csv")
     snowball(
         input_path=INPUT_DIR / "snowballing_openalex.csv",
         output_path=out_fp,
         forward=False,
         backward=True,
         use_all=False,
+        email=EMAIL,
     )
     df = pd.read_csv(out_fp)
-    assert len(df) == 31
+    assert len(df) == 40
 
     all_out_fp = Path(tmpdir, "backward_all.csv")
     snowball(
@@ -99,9 +131,10 @@ def test_openalex_id_backward(tmpdir):
         forward=False,
         backward=True,
         use_all=True,
+        email=EMAIL,
     )
     df_all = pd.read_csv(all_out_fp)
-    assert len(df_all) == 117
+    assert len(df_all) == 45
 
 
 def test_snowballing_from_doi(tmpdir):
@@ -112,6 +145,7 @@ def test_snowballing_from_doi(tmpdir):
         forward=False,
         backward=True,
         use_all=True,
+        email=EMAIL,
     )
     df = pd.read_csv(out_fp)
-    assert len(df) == 117
+    assert len(df) == 45