Merge pull request #78 from jjjermiah/development

Development

.github/workflows/main.yml

@@ -6,7 +6,7 @@ on:
     # push to any branch
     branches: [ main ]
   pull_request:
-    branches: [ main ]
+    branches: [ main , development]
 
 jobs:  
   Continuous-Integration:
@@ -77,7 +77,7 @@ jobs:
     # needs: Continuous-Integration
 
     # if pulling to main, deploy to PyPI
-    if: github.ref == 'refs/heads/main'
+    # if: github.ref == 'refs/heads/main'
 
     # Set up operating system
     runs-on: ubuntu-latest
@@ -257,12 +257,21 @@ jobs:
   Update-README:
     needs: Continuous-Deployment
     runs-on: ubuntu-latest
+    # if: github.ref == 'refs/heads/development'
     steps:
       - name: Checkout code
         uses: actions/checkout@v3
 
-      - name: Pull latest changes
-        run: git pull
+      - name: Get Current branch
+        run: |
+          echo "Current branch is: ${{ github.ref }}"
+          echo "Current branch is: ${{ github.head_ref }}"
+          echo "Current branch is: ${{ github.base_ref }}"
+          echo "Current branch is: ${{ github.event_name }}"
+
+          # fix diverged branch
+          git fetch origin ${{ github.head_ref }}
+          git checkout ${{ github.head_ref }}
 
       - name: Set up Python ${{ matrix.python-version }}
         uses: actions/setup-python@v4
@@ -271,7 +280,11 @@ jobs:
 
       - name: Install using PyPi
         run: |
-          pip install  nbiatoolkit;
+          # update pip
+          pip install --upgrade pip
+
+          # make sure pypi packages are up to date
+          pip install --upgrade nbiatoolkit
           NBIAToolkit
 
       - name: Update README code block
@@ -306,10 +319,10 @@ jobs:
               git add .
 
               # Commit with a timestamp
-              git commit -m "Auto commit: $(date)"
+              git commit -m "Update README with latest version: $LATEST_TAG"
 
               # Push changes to the remote repository
-              git push origin main
+              git push origin HEAD:${{ github.head_ref }}
 
               echo "Changes committed and pushed successfully."
           else

README.md

@@ -50,7 +50,7 @@ nbia-toolkit also provides a command line interface (CLI) to query the NBIA data
 /_/ |_/_____/___/_/  |_/_/  \____/\____/_/_/|_/_/\__/  
 
 
-Version: 0.19.1
+Version: 0.19.2
 
 Available CLI tools: 
 

docs/Tutorial.ipynb

@@ -23,7 +23,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": 2,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -393,6 +393,77 @@
     "pprint(patients[0:5])"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### getNewPatients In Collection\n",
+    "\n",
+    "``` python\n",
+    "getNewPatients(\n",
+    "    Collection: str,   # (required)\n",
+    "    Date: str,      # (required) accepted formats:\n",
+    "                    # \"%Y-%m-%d\", \"%Y/%m/%d\", \"%Y%m%d\", \n",
+    "                    # \"%m/%d/%Y\", \"%d/%m/%Y\", \"%d-%m-%Y\"\n",
+    ")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total new patients in TCGA-BLCA after 2019-01-01: 15\n",
+      "[{'Collection': 'TCGA-BLCA',\n",
+      "  'PatientId': 'TCGA-4Z-AA86',\n",
+      "  'PatientName': 'TCGA-4Z-AA86',\n",
+      "  'PatientSex': 'M',\n",
+      "  'Phantom': 'NO',\n",
+      "  'SpeciesCode': '337915000',\n",
+      "  'SpeciesDescription': 'Homo sapiens'},\n",
+      " {'Collection': 'TCGA-BLCA',\n",
+      "  'PatientId': 'TCGA-G2-A2EC',\n",
+      "  'PatientName': 'TCGA-G2-A2EC',\n",
+      "  'PatientSex': 'F',\n",
+      "  'Phantom': 'NO',\n",
+      "  'SpeciesCode': '337915000',\n",
+      "  'SpeciesDescription': 'Homo sapiens'},\n",
+      " {'Collection': 'TCGA-BLCA',\n",
+      "  'EthnicGroup': 'W',\n",
+      "  'PatientId': 'TCGA-G2-A2EF',\n",
+      "  'PatientName': 'TCGA-G2-A2EF',\n",
+      "  'PatientSex': 'M',\n",
+      "  'Phantom': 'NO',\n",
+      "  'SpeciesCode': '337915000',\n",
+      "  'SpeciesDescription': 'Homo sapiens'},\n",
+      " {'Collection': 'TCGA-BLCA',\n",
+      "  'PatientId': 'TCGA-G2-A2EJ',\n",
+      "  'PatientName': 'TCGA-G2-A2EJ',\n",
+      "  'PatientSex': 'F',\n",
+      "  'Phantom': 'NO',\n",
+      "  'SpeciesCode': '337915000',\n",
+      "  'SpeciesDescription': 'Homo sapiens'},\n",
+      " {'Collection': 'TCGA-BLCA',\n",
+      "  'EthnicGroup': 'W',\n",
+      "  'PatientId': 'TCGA-G2-A2EK',\n",
+      "  'PatientName': 'TCGA-G2-A2EK',\n",
+      "  'PatientSex': 'M',\n",
+      "  'Phantom': 'NO',\n",
+      "  'SpeciesCode': '337915000',\n",
+      "  'SpeciesDescription': 'Homo sapiens'}]\n"
+     ]
+    }
+   ],
+   "source": [
+    "newPatients = client.getNewPatients(Collection=\"TCGA-BLCA\", Date=\"2019-01-01\")\n",
+    "print(f\"Total new patients in TCGA-BLCA after 2019-01-01: {len(newPatients)}\")\n",
+    "pprint(newPatients[0:5])"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {},

src/nbiatoolkit/nbia.py

@@ -1,6 +1,6 @@
 from .auth import OAuth2
 from .logger.logger import setup_logger
-from .utils import NBIA_ENDPOINTS, validateMD5, clean_html, convertMillis
+from .utils import NBIA_ENDPOINTS, validateMD5, clean_html, convertMillis, convertDateFormat
 from .dicomsort import DICOMSorter
 
 import requests
@@ -172,6 +172,34 @@ def getPatients(self, Collection: str = "") -> Union[list[dict[str, str]], None]
 
         return patientList
 
+    def getNewPatients(self, Collection: str, Date: str) -> Union[list[dict[str, str]], None]:
+        assert Collection is not None
+        assert Date is not None
+
+        # convert date to %Y/%m/%d format
+        Date = convertDateFormat(input_date=Date, format="%Y/%m/%d")
+
+        PARAMS = self.parsePARAMS(locals())
+
+        response = self.query_api(endpoint=NBIA_ENDPOINTS.GET_NEW_PATIENTS_IN_COLLECTION, params=PARAMS)
+        assert isinstance(response, list), "Expected list, but received: %s" % type(response)
+
+        patientList = []
+        for patient in response:
+            assert isinstance(patient, dict), "Expected dict, but received: %s" % type(
+                patient
+            )
+            assert "PatientId" in patient, "PatientId not in patient dict"
+            assert isinstance(
+                patient["PatientId"], str
+            ), "PatientId must be a string, but received: %s" % type(
+                patient["PatientId"]
+            )
+
+            patientList.append(patient)
+
+        return patientList
+
     def getPatientsByCollectionAndModality(
         self, Collection: str, Modality: str
     ) -> Union[list[str], None]:

src/nbiatoolkit/utils/__init__.py

@@ -1,4 +1,4 @@
 from .nbia_endpoints import NBIA_ENDPOINTS
 from .md5 import validateMD5
-from .parsers import convertMillis, clean_html
-__all__ = ["NBIA_ENDPOINTS", "validateMD5", "convertMillis", "clean_html"]
+from .parsers import convertMillis, clean_html, convertDateFormat
+__all__ = ["NBIA_ENDPOINTS", "validateMD5", "convertMillis", "clean_html", "convertDateFormat"]

src/nbiatoolkit/utils/nbia_endpoints.py

@@ -15,9 +15,9 @@ class NBIA_ENDPOINTS(Enum):
     GET_MODALITY_VALUES = "v2/getModalityValues"
 
     GET_PATIENTS = "v2/getPatient"
+    GET_NEW_PATIENTS_IN_COLLECTION = "v2/NewPatientsInCollection"
     GET_PATIENT_BY_COLLECTION_AND_MODALITY = "v2/getPatientByCollectionAndModality"
     GET_BODY_PART_PATIENT_COUNT = "getBodyPartValuesAndCounts"
-    GET_NEW_PATIENTS_IN_COLLECTION = "NewPatientsInCollection"
 
     GET_STUDIES = "v2/getPatientStudy"
 

src/nbiatoolkit/utils/parsers.py

@@ -36,4 +36,33 @@ def convertMillis(millis: int) -> str:
         AssertionError: If the input is not an integer.
     """
     assert isinstance(millis, int), "The input must be an integer"
-    return datetime.fromtimestamp(millis / 1000.0).strftime('%Y-%m-%d')
+    return datetime.fromtimestamp(millis / 1000.0).strftime('%Y-%m-%d')
+
+
+def convertDateFormat(input_date: str, format: str = "%Y/%m/%d") -> str:
+    """
+    Converts the input date to the desired format.
+
+    Args:
+        input_date (str): The date to be converted.
+
+    Returns:
+        str: The converted date in the format "YYYY/MM/DD".
+
+    Raises:
+        ValueError: If the input date has an invalid format.
+    """
+    # List of possible date formats with only days, months, and years
+    possible_formats = [
+        "%Y-%m-%d", "%Y/%m/%d", "%Y%m%d", "%m/%d/%Y", "%d/%m/%Y", "%d-%m-%Y"
+    ]
+    # Try parsing the input date with each possible format
+    for date_format in possible_formats:
+        try:
+            parsed_date = datetime.strptime(input_date, date_format)
+            # If parsing is successful, format the date in YYYY/MM/DD and return
+            return parsed_date.strftime(format)
+        except ValueError:
+            pass  # If parsing fails, continue with the next format
+    # If none of the formats match, raise an exception or return a default value
+    raise ValueError("Invalid date format: {}".format(input_date))

tests/test_nbia.py

@@ -92,6 +92,22 @@ def test_getPatients(nbia_patients):
     assert "Collection" in nbia_patients[0]
     assert "PatientSex" in nbia_patients[0]
 
+def test_getNewPatients(nbia_client):
+    patients = nbia_client.getNewPatients('TCGA-BLCA', Date = "2019-01-01")
+    assert isinstance(patients, list)
+    assert len(patients) > 0
+    assert isinstance(patients[0], dict)
+    assert "PatientId" in patients[0]
+    assert "PatientName" in patients[0]
+    assert "Collection" in patients[0]
+    assert "PatientSex" in patients[0]
+
+def test_failed_getNewPatients(nbia_client):
+    with pytest.raises(Exception):
+        patients = nbia_client.getNewPatients('TCGA-BLCA', Date = "bad_date")
+    with pytest.raises(Exception):
+        patients = nbia_client.getNewPatients('bad_collection', Date = "2019-01-01")
+
 def test_getPatientsByCollectionAndModality(nbia_patientsByCollectionAndModality):
     assert isinstance(nbia_patientsByCollectionAndModality, list)
     assert len(nbia_patientsByCollectionAndModality) > 0

tests/test_parsers.py

@@ -1,6 +1,8 @@
-from src.nbiatoolkit.utils.parsers import clean_html, convertMillis
+from src.nbiatoolkit.utils.parsers import clean_html, convertMillis, convertDateFormat
 from datetime import datetime
 import pytest
+
+
 def test_clean_html_valid_input():
     # Test case for valid input with HTML tags and special characters
     html_string = "<p>This is <b>bold</b> text with special characters: &amp; &lt; &gt;</p>"
@@ -41,4 +43,24 @@ def test_convertMillis_invalid_input():
         convertMillis(millis)  # type: ignore
         assert False, "Expected AssertionError"
     except AssertionError as e:
-        assert str(e) == "The input must be an integer"
+        assert str(e) == "The input must be an integer"
+
+
+def test_convertDateFormat_valid_input():
+    # Test case for valid input date in different formats
+    input_date = "2021-09-01"
+    expected_output = "2021/09/01"
+    assert convertDateFormat(input_date) == expected_output
+
+def test_convertDateFormat_invalid_input():
+    # Test case for invalid input date format
+    input_date = "shoulD_fail"  # Invalid format: day-month-year
+    with pytest.raises(ValueError) as e:
+        convertDateFormat(input_date)
+    assert str(e.value) == "Invalid date format: {}".format(input_date)
+
+def test_convertDateFormat_default_value():
+    # Test case for input date with no matching format
+    input_date = "2021/09/01"
+    default_value = "N/A"
+    assert convertDateFormat(input_date, default_value) == default_value