Upload folder using huggingface_hub

graph.py

@@ -519,9 +519,14 @@ async def fn_interpreter_node(state: SheamiState):
     get_db().update_run_stats(run_id=state["run_id"], status="completed")
 
     ## add parsed reports
-    get_db().add_report_v2(
+    report_id_list = get_db().add_report_v2(
         patient_id=state["patient_id"], reports=state["standardized_reports"]
     )
+    state["report_id_list"] = report_id_list
+
+    logger.info("report_id_list = %s", report_id_list)
+    for report_id in report_id_list.split(","):
+        get_db().aggregate_trends_from_report(state["patient_id"], report_id)
 
     return state
 

home.py

@@ -53,10 +53,45 @@ def flatten_reports_v2(reports: List[Dict[str, Any]]) -> pd.DataFrame:
         lab_results = parsed_v2.get("lab_results", [])
 
         if not lab_results:
-            rows.append({
-                "report_id": rid,
-                "uploaded_at": uploaded_at,
-                "file_name": file_name,
+            rows.append(
+                {
+                    "report_id": rid,
+                    "uploaded_at": uploaded_at,
+                    "file_name": file_name,
+                    "test_name": "",
+                    "result_value": "",
+                    "test_unit": "",
+                    "ref_min": "",
+                    "ref_max": "",
+                    "ref_raw": "",
+                    "test_date": "",
+                    "inferred_range": "",
+                }
+            )
+        else:
+            for t in lab_results:
+                ref = t.get("test_reference_range", {})
+                rows.append(
+                    {
+                        "report_id": rid,
+                        "uploaded_at": uploaded_at,
+                        "file_name": file_name,
+                        "test_name": t.get("test_name", ""),
+                        "result_value": t.get("result_value", ""),
+                        "test_unit": t.get("test_unit", ""),
+                        "ref_min": ref.get("min", "-"),
+                        "ref_max": ref.get("max", "-"),
+                        "ref_raw": ref.get("raw", "-"),
+                        "test_date": t.get("test_date", ""),
+                        "inferred_range": t.get("inferred_range", ""),
+                    }
+                )
+    if not rows:
+        rows = [
+            {
+                "report_id": "",
+                "uploaded_at": "",
+                "file_name": "",
                 "test_name": "",
                 "result_value": "",
                 "test_unit": "",
@@ -65,39 +100,10 @@ def flatten_reports_v2(reports: List[Dict[str, Any]]) -> pd.DataFrame:
                 "ref_raw": "",
                 "test_date": "",
                 "inferred_range": "",
-            })
-        else:
-            for t in lab_results:
-                ref = t.get("test_reference_range", {})
-                rows.append({
-                    "report_id": rid,
-                    "uploaded_at": uploaded_at,
-                    "file_name": file_name,
-                    "test_name": t.get("test_name", ""),
-                    "result_value": t.get("result_value", ""),
-                    "test_unit": t.get("test_unit", ""),
-                    "ref_min": ref.get("min","-"),
-                    "ref_max": ref.get("max","-"),
-                    "ref_raw": ref.get("raw", "-"),
-                    "test_date": t.get("test_date", ""),
-                    "inferred_range": t.get("inferred_range", ""),
-                })
-    if not rows:
-        rows = [{
-            "report_id": "",
-            "uploaded_at": "",
-            "file_name": "",
-            "test_name": "",
-            "result_value": "",
-            "test_unit": "",
-            "ref_min": "",
-            "ref_max": "",
-            "ref_raw": "",
-            "test_date": "",
-            "inferred_range": "",
-        }]
+            }
+        ]
     df = pd.DataFrame(rows)
-    df = df.fillna("-")   # 👈 normalize missing values
+    df = df.fillna("-")  # 👈 normalize missing values
     return df
 
 
@@ -370,9 +376,13 @@ with gr.Blocks(
         user_email_state = gr.State("")
         patient_id_state = gr.State("")
         # show them if you want
-        shown_email = gr.Textbox(label="User Email", interactive=False)
-        shown_patient = gr.Textbox(label="Patient ID", interactive=False)        
-    get_gradio_block(container=upload_reports_modal,user_email_state=user_email_state, patient_id_state=patient_id_state)
+        shown_email = gr.Textbox(label="User Email", interactive=False, visible=False)
+        shown_patient = gr.Textbox(label="Patient ID", interactive=False, visible=False)
+    get_gradio_block(
+        container=upload_reports_modal,
+        user_email_state=user_email_state,
+        patient_id_state=patient_id_state,
+    )
 
     with gr.Sidebar() as sheami_sidebar:  # Sidebar
         # gr.Markdown("### Sidebar")
@@ -514,18 +524,24 @@ with gr.Blocks(
     # open modal and set states
     def show_upload_reports_modal(user_email, patient_id):
         return [
-            user_email, 
-            patient_id, 
-            gr.update(value=user_email), 
-            gr.update(value=patient_id), 
+            user_email,
+            patient_id,
+            gr.update(value=user_email),
+            gr.update(value=patient_id),
             gr.update(visible=True),
         ]
 
     upload_reports_btn.click(close_side_bar, outputs=[sheami_sidebar]).then(
         show_upload_reports_modal,
         inputs=[email_in, patient_list],
-        outputs=[user_email_state, patient_id_state, shown_email, shown_patient, upload_reports_modal],
-    )    
+        outputs=[
+            user_email_state,
+            patient_id_state,
+            shown_email,
+            shown_patient,
+            upload_reports_modal,
+        ],
+    )
 
 if __name__ == "__main__":
     demo.launch()

modules/db.py

@@ -7,6 +7,7 @@ from dotenv import load_dotenv
 
 from modules.models import StandardizedReport
 
+
 class SheamiDB:
     def __init__(self, uri: str, db_name: str = "sheami"):
         """Initialize connection to MongoDB Atlas (or local Mongo)."""
@@ -70,7 +71,7 @@ class SheamiDB:
     # REPORT FUNCTIONS
     # ---------------------------
     def add_report_v2(self, patient_id: str, reports: list[StandardizedReport]) -> str:
-        inserted_ids = []
+        inserted_ids: list[ObjectId] = []
         for parsed_data in reports:
             report = {
                 "patient_id": ObjectId(patient_id),
@@ -80,7 +81,7 @@ class SheamiDB:
             }
             result = self.reports.insert_one(report)
             inserted_ids.append(result.inserted_id)
-        return str(inserted_ids)
+        return ",".join([str(inserted_id) for inserted_id in inserted_ids])
 
     def add_report(self, patient_id: str, file_name: str, parsed_data: any) -> str:
         report = {
@@ -305,3 +306,69 @@ class SheamiDB:
             )
 
         return result.modified_count
+
+    def aggregate_trends_from_report(self, patient_id: str, report_id: str):
+        """
+        Incrementally update patient trends based on a new report's tests.
+        - Fetches the report
+        - For each test, appends (date, value, unit) to trends[patient_id, test_name]
+        - Ensures no duplicate points for same report/test combo
+        """
+        report = self.reports.find_one({"_id": ObjectId(report_id)})
+        if not report:
+            raise ValueError(f"Report {report_id} not found")
+
+        # print("report = ",report)
+        tests = report.get("parsed_data_v2", {"lab_results": []}).get("lab_results", [])
+        if not tests:
+            return 0
+
+        updated = 0
+        for test in tests:
+            test_name = test.get("test_name")
+            if not test_name:
+                continue
+
+            value = test.get("result_value")
+            unit = test.get("test_unit")
+            test_date = (
+                test.get("test_date")
+                or report.get("uploaded_at")
+                or datetime.now(timezone.utc)
+            )
+
+            # Normalize date
+            if isinstance(test_date, (int, float)):
+                # handle timestamp
+                test_date = datetime.fromtimestamp(test_date, tz=timezone.utc)
+            elif isinstance(test_date, str):
+                try:
+                    test_date = datetime.fromisoformat(test_date)
+                except Exception:
+                    test_date = datetime.now(timezone.utc)
+
+            point = {
+                "date": test_date,
+                "value": value,
+                "unit": unit,
+                "report_id": ObjectId(report_id),
+            }
+
+            # Upsert trend doc
+            result = self.trends.update_one(
+                {"patient_id": ObjectId(patient_id), "test_name": test_name},
+                {
+                    "$setOnInsert": {
+                        "patient_id": ObjectId(patient_id),
+                        "test_name": test_name,
+                        "created_at": datetime.now(timezone.utc),
+                    },
+                    "$push": {"trend_data": point},
+                    "$set": {"last_updated": datetime.now(timezone.utc)},
+                },
+                upsert=True,
+            )
+            updated += result.modified_count
+
+        # print("updated/inserted", updated, "trends")
+        return updated

modules/models.py

@@ -3,6 +3,7 @@ from datetime import datetime
 from pydantic import BaseModel, Field
 from typing import List, Literal, Optional, TypedDict, Union
 
+
 class PatientInfo(BaseModel):
     name: Optional[str] = Field(None, description="Patient's full name")
     age: Optional[int] = Field(None, description="Patient's age in years")
@@ -84,6 +85,7 @@ class SheamiMilestone:
 class SheamiState(TypedDict):
     user_email: str
     patient_id: str
+    report_id_list: str
     run_id: str
     messages: list[str]
     thread_id: str