UnboundLocalError with fuzzy matching

[SLL-26]

Description

Calling find_drugs() with is_fuzzy_match=True raises an UnboundLocalError instead of returning drug matches.

The error occurs when processing a tokenised sentence containing medication names and dosages. It appears that the variable lookup_name is referenced before being assigned within the package's find_drugs() implementation.

The issue occurs consistently with the example below and prevents fuzzy matching from being used. The same code works correctly when fuzzy matching is disabled (is_fuzzy_match=False).

Environment

• Operating System: Windows
• Python Version: 3.13.7
• spaCy Version: 3.8.15
• drug_named_entity_recognition Version: 2.0.9
• Environment: Python virtual environment (venv)

How to Reproduce

1. Create a spaCy document:

test = "Repeat prescription list still shows Mirtazapine 45mg Nocte and paracetomol 500mg OD"
test_doc = nlp(test)

2. Run find_drugs() with fuzzy matching enabled:

matches = find_drugs(
    [t.text for t in test_doc],
    is_fuzzy_match=True
)

3. Observe the exception:

UnboundLocalError: cannot access local variable 'lookup_name'
where it is not associated with a value

Expected Behaviour

find_drugs() should return drug matches when fuzzy matching is enabled. For example, it should identify:

• Mirtazapine
• paracetomol (via fuzzy matching to paracetamol, if supported)

Actual Behaviour

The function raises:

UnboundLocalError: cannot access local variable 'lookup_name'
where it is not associated with a value

Full Traceback

---------------------------------------------------------------------------
UnboundLocalError                         Traceback (most recent call last)

matches = find_drugs([t.text for t in test_doc], is_fuzzy_match=True)

File ".../drug_named_entity_recognition/drugs_finder.py", line 268, in find_drugs
    match_data = get_molecular_weight(
        match_data, lookup_name, use_pub_chem_api
    )

UnboundLocalError: cannot access local variable 'lookup_name'
where it is not associated with a value

Additional Information

The issue does not occur when fuzzy matching is disabled:

matches = find_drugs(
    [t.text for t in test_doc],
    is_fuzzy_match=False
)

which returns results as expected.

[SLL-26]

Additional Debugging

The error can be reproduced with the following code:

test = "Repeat prescription list still shows Mirtazapine 45mg Nocte and paracetomol 500mg OD"
test_doc = nlp(test)

tokens = [t.text for t in test_doc]

for i in range(1, len(tokens) + 1):
    try:
        find_drugs(tokens[:i], is_fuzzy_match=True)
        print(i, "OK")
    except Exception as e:
        print(i, "FAIL", e)
        break

Output:

1 OK
2 OK
3 OK
4 OK
5 OK
6 OK
7 OK
8 OK
9 FAIL cannot access local variable 'lookup_name' where it is not associated with a value

The error first occurs when processing the first 9 tokens of the sentence.

print(tokens[:9])

Output:

[
    'Repeat',
    'prescription',
    'list',
    'still',
    'shows',
    'Mirtazapine',
    '45',
    'mg',
    'Nocte'
]

Notably, the error occurs before the misspelled token "paracetomol" is encountered, suggesting the issue is probably unrelated to fuzzy matching of the second drug mention.