--- a
+++ b/LLM-Zero-shot_approach/utils.py
@@ -0,0 +1,62 @@
+def build_prompt(sentence, index_propmt):
+    """
+    Given a sentence and the index of the prompt, it returns the prompt to be used for NER.
+    args:
+    sentence: str, sentence to be annotated
+    index_propmt: int, index of the prompt to be used
+    return:
+    prompt: str, prompt to be used for NER
+    """
+    prompt_1 = '''I need to perform a named entity recognition task on a  text related with inclusion criteria in clinical trials.
+    The entities you need to recognize are: Condition, Value, Drug, Procedure, Measurement, Temporal, Observation, Person, Mood, Device and Pregnancy_considerations.
+    Particularly you have to produce the ouput in the BIO format. I will show you an example of the expected output.
+    Input text: Patients with symptomatic CNS metastases or leptomeningeal involvement
+    Output:
+    Patients O
+    with O
+    symptomatic O
+    CNS B-Condition
+    metastases I-Condition
+    or O
+    leptomeningeal B-Condition
+    involvement I-Condition
+
+    You can see that tokens without any entity are labeled as O, and the tokens that are part of an entity are labeled as B-<entity> or I-<entity> depending on if they are the beginning or the inside of the entity.
+    Please, just answer the question for this specific example and stop writting after that.
+    Input text: '''
+    promt_2 = '''I am working on a named entity recognition problem, in the context of clinical
+    trials eligibility criteria. I will show you the list of entities:
+    - Condition
+    - Value
+    - Drug
+    - Procedure
+    - Measurement
+    - Temporal
+    - Observation
+    - Person
+    - Mood
+    - Device
+
+    Your task consists in annotate the named entities in a given sentence in the format I will explain you.
+    I will explain you with some examples:
+
+    Example 1:
+    Input: Patients who have received prior chemotherapy for unresectable disease.
+    Output: Patients who have received prior <Procedure>chemotherapy</Procedure> for <Condition>unresectable disease</Condition>.
+
+    Example 2:
+    Input: Patients with any other severe concurrent disease, which in the judgment of the investigator, would make the patient inappropriate for entry into this study.
+    Ouput: Patients with any other severe <Condition>concurrent disease</Condition>, which in the judgment of the investigator, would make the patient <Mood>inappropriate for <Observation>entry into this study</Observation>.
+
+    As you can see, in each example, the extracted entities are enclosed using the sintax: <ENT>text of the entity</ENT>.
+
+    Please now annotate as explained before the following sentence:
+
+    Input: '''
+
+    if index_propmt == 1:
+        prompt = prompt_1 + sentence
+    else:
+        prompt = promt_2 + sentence
+    return prompt
+