Spaces:

Fuzure
/

sheami

Sleeping

Upload folder using huggingface_hub
db5222a verified about 1 year ago
30.7 kB
from datetime import timezone, date
import os
import random
from typing import Any
from datetime import datetime, timezone
from bson import ObjectId
from dotenv import load_dotenv
from modules.models import StandardizedReport
from motor.motor_asyncio import AsyncIOMotorGridFSBucket
from motor.motor_asyncio import AsyncIOMotorClient, AsyncIOMotorGridFSBucket
import os
from dotenv import load_dotenv
import pandas as pd

class SheamiDB:
    def __init__(self, uri: str = None, db_name: str = "sheami"):
        if not uri:
            load_dotenv(override=True)
            uri = os.getenv("MONGODB_URI")

        # Use Motor's AsyncIOMotorClient instead of PyMongo's AsyncMongoClient
        self.client = AsyncIOMotorClient(uri)
        self.db = self.client[db_name]

        # Collections
        self.users = self.db["users"]
        self.patients = self.db["patients"]
        self.reports = self.db["reports"]
        self.trends = self.db["trends"]
        self.final_reports = self.db["final_reports"]
        self.run_stats = self.db["run_stats"]
        self.vitals = self.db["vitals"]

        # Motor's GridFSBucket requires a MotorDatabase
        self.fs = AsyncIOMotorGridFSBucket(self.db)

    # ---------------------------
    # USER FUNCTIONS
    # ---------------------------
    async def add_user(self, email: str, name: str) -> str:
        user = {"email": email, "name": name, "created_at": datetime.now(timezone.utc)}
        result = await self.users.insert_one(user)
        return str(result.inserted_id)

    async def get_user(self, user_id: str) -> dict:
        user = await self.users.find_one({"_id": ObjectId(user_id)})
        return user

    def convert_dob(self, dob: int | float | datetime):
        if isinstance(dob, datetime):
            transformed_dob = dob  # PyMongo will store it as ISODate automatically
        elif isinstance(dob, (int, float)):
            transformed_dob = datetime.fromtimestamp(dob)
        else:
            transformed_dob = dob
        return transformed_dob

    # ---------------------------
    # PATIENT FUNCTIONS
    # ---------------------------
    async def add_patient(
        self, user_id: str, name: str, dob: int | float | datetime, gender: str
    ) -> str:
        # Ensure DOB is stored as MongoDB Date
        transformed_dob = self.convert_dob(dob)
        patient = {
            "user_id": ObjectId(user_id),
            "name": name,
            "dob": transformed_dob,
            "gender": gender,
            "created_at": datetime.now(timezone.utc),
        }
        result = await self.patients.insert_one(patient)
        return str(result.inserted_id)

    async def get_patient_by_id(
        self, patient_id: str, fields: list[str] = [], serializable: bool = False
    ) -> Any | None:
        patient = await self.patients.find_one({"_id": ObjectId(patient_id)})
        if fields:
            patient = {key: patient[key] for key in fields if key in patient}
        if serializable:
            patient = self.convert_to_serializable_data(data=patient)
        return patient

    async def get_patients_by_user(self, user_id: str) -> list:
        cursor = self.patients.find({"user_id": ObjectId(user_id)}).sort("name")
        # all the locked records are example records
        example_cursor = self.patients.find(
            {"user_id": ObjectId("68a40aa2208e5689f7342a6e"), "locked": True}
        ).sort("name")
        matching_patients = await cursor.to_list(
            length=None
        )  # length=None returns all documents

        example_patients = await example_cursor.to_list(length=None)

        sorted_list = sorted(
            example_patients + matching_patients, key=lambda x: x["name"]
        )

        return sorted_list

    # ---------------------------
    # REPORT FUNCTIONS
    # ---------------------------
    async def add_report_v2(
        self, patient_id: str, reports: list[StandardizedReport], run_id: str
    ) -> str:
        inserted_ids: list[ObjectId] = []
        for parsed_data in reports:
            report = {
                "patient_id": ObjectId(patient_id),
                "uploaded_at": datetime.now(timezone.utc),
                "file_name": parsed_data.original_report_file_name,
                "parsed_data_v2": parsed_data.model_dump(),
                "run_id": ObjectId(run_id),
            }
            result = await self.reports.insert_one(report)
            inserted_ids.append(result.inserted_id)
        return ",".join([str(inserted_id) for inserted_id in inserted_ids])

    async def add_report(
        self, patient_id: str, file_name: str, parsed_data: any
    ) -> str:
        report = {
            "patient_id": ObjectId(patient_id),
            "uploaded_at": datetime.now(timezone.utc),
            "file_name": file_name,
            "parsed_data": parsed_data,
        }
        result = await self.reports.insert_one(report)
        return str(result.inserted_id)

    async def get_reports_by_patient(self, patient_id: str) -> list:
        reports_cursor = self.reports.find({"patient_id": ObjectId(patient_id)}).sort(
            "_id", -1
        )
        reports = await reports_cursor.to_list(
            length=None
        )  # Fetch all sorted documents as a list
        return reports

    # ---------------------------
    # TREND FUNCTIONS
    # ---------------------------
    async def add_or_update_trend(
        self, patient_id: str, test_name: str, trend_data: list
    ):
        """Insert new trend or update existing one."""
        await self.trends.update_one(
            {"patient_id": ObjectId(patient_id), "test_name": test_name},
            {
                "$set": {
                    "trend_data": trend_data,
                    "last_updated": datetime.now(timezone.utc),
                }
            },
            upsert=True,
        )

    async def get_trends_by_patient(
        self, patient_id: str, fields: list[str] = None, serializable=False
    ) -> list:
        cursor = self.trends.find({"patient_id": ObjectId(patient_id)}).sort(
            "test_name"
        )
        trends = await cursor.to_list(length=None)
        if fields:
            trends = [
                {field: trend[field] for field in fields if field in trend}
                for trend in trends
            ]
        if serializable:
            trends = self.convert_to_serializable_data(data=trends)

        return trends

    # ---------------------------
    # FINAL REPORT FUNCTIONS
    # ---------------------------
    async def add_final_report(
        self,
        patient_id: str,
        summary: str,
        recommendations: list,
        trend_snapshots: list,
    ) -> str:
        final_report = {
            "patient_id": ObjectId(patient_id),
            "generated_at": datetime.now(timezone.utc),
            "summary": summary,
            "recommendations": recommendations,
            "trend_snapshots": trend_snapshots,
        }
        result = await self.final_reports.insert_one(final_report)
        return str(result.inserted_id)

    async def get_final_reports_by_patient(self, patient_id: str) -> list:
        cursor = self.final_reports.find({"patient_id": ObjectId(patient_id)}).sort(
            "_id", -1
        )
        final_reports = await cursor.to_list(length=None)
        return final_reports

    # ---------------------------
    # FETCH FULL USER DATA
    # ---------------------------
    async def get_user_by_email(self, email: str) -> dict:
        """Fetch user by email."""
        user = await self.users.find_one({"email": email})
        return user

    async def get_user_full_data(self, user_id: str) -> dict:
        """
        Fetch user + all patients, reports, trends, final reports
        for populating UI (tabbed layout).
        """
        user = await self.get_user(user_id)
        if not user:
            return {}

        # Get patients for user
        patients = await self.get_patients_by_user(user_id)
        full_patients = []

        for patient in patients:
            pid = str(patient["_id"])

            # Fetch related collections
            patient_reports = await self.get_reports_by_patient(pid)
            patient_trends = await self.get_trends_by_patient(pid)
            patient_final_reports = await self.get_final_reports_by_patient(pid)

            full_patients.append(
                {
                    "patient": patient,
                    "reports": patient_reports,
                    "trends": patient_trends,
                    "final_reports": patient_final_reports,
                }
            )

        return {"user": user, "patients": full_patients}

    async def update_patient(self, patient_id, fields: dict):
        fields["dob"] = self.convert_dob(fields["dob"])
        result = await self.patients.update_one(
            {"_id": ObjectId(patient_id)}, {"$set": fields}
        )
        return result.modified_count > 0

    async def delete_patient(self, patient_id: str):
        try:
            yield "⌛Deleting patient PDFs ... "
            deleted_count = await self.delete_pdfs_by_patient_id(patient_id)
            yield f"✅Deleted {deleted_count} patient PDFs ... "
            yield "⌛Deleting patient reports ... "
            result = await self.reports.delete_one({"patient_id": ObjectId(patient_id)})
            yield f"✅Deleted {result.deleted_count} patient reports ... "
            yield "⌛Deleting patient trends ... "
            result = await self.trends.delete_one({"patient_id": ObjectId(patient_id)})
            yield f"✅Deleted {result.deleted_count} patient trends ... "
            yield "⌛Deleting patient final reports ... "
            result = await self.final_reports.delete_one(
                {"patient_id": ObjectId(patient_id)}
            )
            yield f"✅Deleted {result.deleted_count} patient final reports ... "
            yield "⌛Deleting patient run stats ... "
            result = await self.run_stats.delete_one(
                {"patient_id": ObjectId(patient_id)}
            )
            yield f"✅Deleted {result.deleted_count} patient run stats ... "
            yield "⌛Deleting patient ... "
            result = await self.patients.delete_one({"_id": ObjectId(patient_id)})
            yield f"✅Deleted {result.deleted_count} patient ... "
        except Exception as e:
            print(f"Error deleting patient: {e}")
            yield f"❌ Error deleting patient {e}"
            return

    async def start_run(
        self,
        user_email: str,
        patient_id: str,
        source_file_names: list[str],
        source_file_contents: list[str],
        action: str = "upload test reports",
    ):
        print("Getting details for user:", user_email, " and patient:", patient_id)
        user = await self.get_user_by_email(user_email)
        if not user:
            raise Exception(f"User {user_email} not found!")

        user_id = user.get("_id")
        if not user_id:
            raise Exception(f"User {user_email} not found!")
        if not patient_id:
            raise Exception("Patient not found!")
        if not source_file_names:
            raise Exception("No source files to process!")
        stat_entry = {
            "user_id": ObjectId(user_id),
            "patient_id": ObjectId(patient_id),
            "source_file_names": source_file_names,
            "source_file_contents": source_file_contents,
            "action": action,
            "status": "inprogress",
            "steps_completed": 0,
            "steps_total": 5,  # max_steps
            "milestones": [
                {
                    "milestone": "Initializing",
                    "start_timestamp": datetime.now(timezone.utc),
                    "end_timestamp": datetime.now(timezone.utc),
                    "status": "completed",
                }
            ],
            "created_at": datetime.now(timezone.utc),
        }
        result = await self.run_stats.insert_one(stat_entry)
        return str(result.inserted_id)

    async def update_run_stats(self, run_id: str, **kwargs):
        """
        Update top-level fields in a run's stats document.
        Example usage:
            update_run_stats(run_id, steps_completed=2, steps_total=5)
        """
        update_fields = {}
        for key, value in kwargs.items():
            if key in [
                "steps_completed",
                "steps_total",
                "action",
                "source_file_names",
                "source_file_contents",
                "status",
                "message",
            ]:
                update_fields[key] = value

        if not update_fields:
            raise ValueError("No valid run-level fields to update")

        result = await self.run_stats.update_one(
            {"_id": ObjectId(run_id)}, {"$set": update_fields}
        )
        return result.modified_count

    async def add_or_update_milestone(
        self, run_id: str, milestone: str, status: str = None, end: bool = False
    ):
        """
        Add a new milestone or update an existing one.
        - If milestone doesn't exist, pushes a new entry with start_timestamp.
        - If exists, updates status or end_timestamp.
        - If milestone completes (end=True), also increments steps_completed at run level.
        """
        run = await self.run_stats.find_one({"_id": ObjectId(run_id)})
        if not run:
            raise Exception(f"Run {run_id} not found")

        milestones = run.get("milestones", [])
        existing = next((m for m in milestones if m["milestone"] == milestone), None)

        if not existing:
            # add new milestone
            new_milestone = {
                "milestone": milestone,
                "start_timestamp": datetime.now(timezone.utc),
                "end_timestamp": None,
                "status": status or "inprogress",
            }
            result = await self.run_stats.update_one(
                {"_id": ObjectId(run_id)}, {"$push": {"milestones": new_milestone}}
            )
        else:
            # update existing milestone
            updates = {}
            if status:
                updates["milestones.$.status"] = status
            if end:
                updates["milestones.$.end_timestamp"] = datetime.now(timezone.utc)

            if not updates:
                raise ValueError("Nothing to update in milestone")

            update_ops = {"$set": updates}
            if end:
                # also increment steps_completed at run level
                update_ops["$inc"] = {"steps_completed": 1}

            result = await self.run_stats.update_one(
                {"_id": ObjectId(run_id), "milestones.milestone": milestone},
                update_ops,
            )

        return result.modified_count

    async 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 = await 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

        async def add_or_update_trend_data_point(test):
            test_name = test.get("test_name")
            value = test.get("result_value")
            unit = test.get("test_unit")
            test_date = test.get("test_date") or datetime.now(timezone.utc)
            test_reference_range = test.get("test_reference_range")
            inferred_range = test.get("inferred_range")

            # Normalize test_date (keep your existing normalization here)...
            if isinstance(test_date, (int, float)):
                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),
            }

            # Step 1: Check if trend_data with same date exists
            existing_doc = await self.trends.find_one(
                {
                    "patient_id": ObjectId(patient_id),
                    "test_name": test_name,
                    "trend_data.date": test_date,
                },
                projection={"trend_data.$": 1},  # Project only matched array element
            )

            if existing_doc:
                # Step 2: Update the existing trend_data array element with new data
                result = await self.trends.update_one(
                    {
                        "patient_id": ObjectId(patient_id),
                        "test_name": test_name,
                        "trend_data.date": test_date,
                    },
                    {
                        "$set": {
                            "trend_data.$.value": value,
                            "trend_data.$.unit": unit,
                            "trend_data.$.report_id": ObjectId(report_id),
                            "last_updated": datetime.now(timezone.utc),
                            "test_reference_range": test_reference_range,
                            "inferred_range": inferred_range,
                        },
                        "$setOnInsert": {
                            "patient_id": ObjectId(patient_id),
                            "test_name": test_name,
                            "created_at": datetime.now(timezone.utc),
                        },
                    },
                )
            else:
                # Step 3: Insert new point as it does not exist yet
                result = await self.trends.update_one(
                    {"patient_id": ObjectId(patient_id), "test_name": test_name},
                    {
                        "$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),
                            "test_reference_range": test_reference_range,
                            "inferred_range": inferred_range,
                            "test_reference_range": test_reference_range,
                            "inferred_range": inferred_range,
                        },
                    },
                    upsert=True,
                )
            return result

        for test in tests:
            test_name = test.get("test_name")
            if not test_name:
                sub_results = test.get("sub_results", [])
                if not sub_results:
                    continue
                for sub_result in sub_results:
                    test_name = sub_result.get("test_name")
                    db_output = await add_or_update_trend_data_point(sub_result)
                    updated += db_output.modified_count
                continue
            else:
                db_output = await add_or_update_trend_data_point(test)
                updated += db_output.modified_count

        # print("updated/inserted", updated, "trends")
        return updated

    async def aggregate_trends_snapshot(self, patient_id: str):
        # fetch trends for this patient
        trend_docs = await self.get_trends_by_patient(patient_id)

        # extract "snapshots"
        snapshots = []
        for t in trend_docs:
            td = t.get("trend_data", [])
            if not td:
                continue
            # last point = most recent measurement
            last_point = max(td, key=lambda x: x.get("date"))
            snapshots.append(
                {
                    "test_name": t.get("test_name", ""),
                    "latest_value": last_point.get("value"),
                    "latest_date": last_point.get("date"),
                    "unit": last_point.get("unit", ""),
                    "reference_range": last_point.get("reference_range", ""),
                }
            )
        return snapshots

    async def upload_bytes_to_fs(self, data: bytes, filename: str, patient_id):
        # Open an upload stream
        upload_stream = self.fs.open_upload_stream(
            filename,
            metadata={
                "patient_id": patient_id,
                "uploaded_at": datetime.now(timezone.utc),
            },
        )
        # Write data to stream
        await upload_stream.write(data)
        await upload_stream.close()
        # The file ID is in upload_stream._id
        return upload_stream._id

    # ---------------------------
    # FINAL REPORT FUNCTIONS
    # ---------------------------
    async def add_final_report_v2(
        self,
        patient_id: str,
        summary: str,
        pdf_bytes: bytes = None,
        file_name: str = None,
    ) -> str:
        """
        Insert a final report.
        If pdf_bytes is provided, stores it in GridFS and saves file_id in metadata.
        """
        pdf_file_id = None
        if pdf_bytes:
            pdf_file_id = await self.upload_bytes_to_fs(
                data=pdf_bytes, filename=file_name, patient_id=ObjectId(patient_id)
            )
        final_report = {
            "patient_id": ObjectId(patient_id),
            "generated_at": datetime.now(timezone.utc),
            "summary": summary,
            "trend_snapshots": await self.aggregate_trends_snapshot(
                patient_id=patient_id
            ),
            "pdf_file_id": pdf_file_id,  # Reference to GridFS file
        }

        result = await self.final_reports.insert_one(final_report)
        return str(result.inserted_id)

    async def get_final_report_pdf(self, report_id: str) -> bytes | None:
        """
        Fetch the PDF bytes for a given final_report (if stored).
        """
        doc = await self.final_reports.find_one({"_id": ObjectId(report_id)})
        if not doc or not doc.get("pdf_file_id"):
            return None
        grid_out = await self.fs.open_download_stream(doc["pdf_file_id"])
        # Read the file data in chunks
        file_data = b""
        while True:
            chunk = await grid_out.read(1024)  # Read 1024 bytes at a time
            if not chunk:
                break
            file_data += chunk
        # await grid_out.close()
        return file_data

    async def get_final_report_html(self, report_id: str) -> bytes | None:
        """
        Fetch the HTML for a given final_report (if stored).
        """
        doc = await self.final_reports.find_one({"_id": ObjectId(report_id)})
        if not doc or not doc.get("summary"):
            return "<html></html>"  # empty tag
        return doc.get("summary")

    async def get_run_stats_by_patient(self, patient_id: str) -> list:
        cursor = self.run_stats.find({"patient_id": ObjectId(patient_id)}).sort(
            "created_at", -1
        )
        run_stats = await cursor.to_list(length=None)
        return run_stats

    async def get_run_stats_by_id(self, id: str):
        run_stat = await self.run_stats.find_one({"_id": ObjectId(id)})
        return run_stat

    async def delete_pdfs_by_patient_id(self, patient_id: str) -> int:
        # Find all files with the specified patient_id in metadata
        cursor = self.db.fs.files.find(
            {"metadata.patient_id": ObjectId(patient_id)}, projection={"_id": 1}
        )
        deleted_count = 0
        async for file_doc in cursor:
            file_id = file_doc["_id"]
            await self.fs.delete(file_id)
            deleted_count += 1
        return deleted_count

    def convert_to_serializable_data(self, data):
        """
        Recursively converts MongoDB-specific types to JSON serializable formats.
        - ObjectId to string
        - datetime to ISO 8601 string
        Handles dict, list, and basic types.
        """
        if isinstance(data, dict):
            return {k: self.convert_to_serializable_data(v) for k, v in data.items()}
        elif isinstance(data, list):
            return [self.convert_to_serializable_data(i) for i in data]
        elif isinstance(data, ObjectId):
            return str(data)
        elif isinstance(data, datetime):
            return data.isoformat()
        else:
            return data

    def normalize_to_date(self, reading_date):
        """
        Convert a datetime or date to a datetime at 00:00:00
        """
        if isinstance(reading_date, date) and not isinstance(reading_date, datetime):
            # convert date to datetime at midnight
            return datetime(reading_date.year, reading_date.month, reading_date.day)
        elif isinstance(reading_date, datetime):
            # strip time part
            return datetime(reading_date.year, reading_date.month, reading_date.day)
        else:
            raise ValueError("reading_date must be a date or datetime")

    async def save_readings_to_db(
        self, patient_id: str, reading_date, new_readings: list, created_by: str
    ):
        """
        Async save/merge readings for a patient/date using Motor.
        """

        reading_date = self.normalize_to_date(reading_date)
        # Find existing document
        doc = await self.vitals.find_one(
            {
                "patient_id": ObjectId(patient_id),
                "date": reading_date,
            }
        )

        if not doc:
            # Insert new
            await self.vitals.insert_one(
                {
                    "patient_id": ObjectId(patient_id),
                    "date": reading_date,
                    "readings": new_readings,
                    "created_by": created_by,
                }
            )
        else:
            # Merge: update existing readings by 'name', append new ones
            existing = doc.get("readings", [])
            merged = existing.copy()

            existing_names = {r["name"]: r for r in existing}
            for r in new_readings:
                if r["name"] in existing_names:
                    # Update existing entry
                    existing_names[r["name"]].update(r)
                else:
                    # Append new entry
                    merged.append(r)

            # For entries that were updated in-place, ensure they are in merged
            merged_names = {r["name"] for r in merged}
            for name, r in existing_names.items():
                if name not in merged_names:
                    merged.append(r)

            await self.vitals.update_one(
                {"_id": doc["_id"]},
                {"$set": {"readings": merged, "created_by": created_by}},
            )

    async def get_vitals_by_patient(
        self, patient_id: str, fields: list[str] = None, serializable=False
    ) -> list:
        cursor = self.vitals.find({"patient_id": ObjectId(patient_id)}).sort("date", -1)
        vitals = await cursor.to_list(length=None)
        if fields:
            vitals = [
                {field: vital[field] for field in fields if field in vital}
                for vital in vitals
            ]
        if serializable:
            vitals = self.convert_to_serializable_data(data=vitals)
        # print("vitals = ",vitals)
        return vitals

    async def get_latest_vitals_by_patient(self, patient_id: str) -> dict:
        # sort by date descending and get the first record
        cursor = self.vitals.find({"patient_id": ObjectId(patient_id)}).sort("date", -1)
        vitals = await cursor.to_list(length=None)
        if len(vitals) > 0:
            vitals = vitals[0]
        else:
            vitals = {}
        return vitals

    async def get_vitals_trends_by_patient(self, patient_id):
        docs = await self.vitals.aggregate([
            {"$match": {"patient_id": ObjectId(patient_id)}},
            {"$unwind": "$readings"},
            {"$project": {
                "date": 1,
                "name": "$readings.name",
                "value": "$readings.value",
                "unit": "$readings.unit"
            }},
            {"$sort": {"date": 1}}
        ]).to_list(length=None)

        df = pd.DataFrame(docs)
        trend_docs = []

        if "name" in df:
            for vital_name, group in df.groupby("name"):
                if vital_name.lower() in ["bp", "blood pressure"]:
                    # Split systolic/diastolic
                    systolic, diastolic = [], []
                    for _, row in group.iterrows():
                        try:
                            sys, dia = row["value"].split("/")
                            systolic.append({"date": row["date"], "value": int(sys)})
                            diastolic.append({"date": row["date"], "value": int(dia)})
                        except Exception:
                            continue

                    # Append as two separate test series
                    trend_docs.append({
                        "test_name": "BP - Systolic",
                        "trend_data": systolic,
                        "unit": "mmHg",
                        "test_reference_range": {}
                    })
                    trend_docs.append({
                        "test_name": "BP - Diastolic",
                        "trend_data": diastolic,
                        "unit": "mmHg",
                        "test_reference_range": {}
                    })
                else:
                    # Other vitals
                    trend_docs.append({
                        "test_name": vital_name,
                        "trend_data": group[["date", "value"]].to_dict("records"),
                        "unit": group["unit"].iloc[0],
                        "test_reference_range": {}
                    })

        # 🔑 Sort by test_name so BP series appear together
        trend_docs = sorted(trend_docs, key=lambda x: x["test_name"])

        return trend_docs