# Cell Throughput Prediction
:::warning
後來覺得預測UE的throughput對於換手沒有幫助，所以預測Cell的throughput來為UE選擇最佳的基地台
:::

## [QP Cell Data](https://gerrit.o-ran-sc.org/r/gitweb?p=ric-app/qp.git;a=blob;f=qp/cell.json.gz;h=bfe57366cbe82b79d824a7e46998b29aa943c2db;hb=48aa41171b8ca141df1f14341ea0ec2fde745af5)


## [SIM E2 Data (including cell data and ue data)](https://github.com/o-ran-sc/sim-e2-interface-data)


throughput_y is the throughput of Cell

throughput_x is the throughput of UE

pdcpBytesDl and availPrbDl has high correlation with throughput of Cell, so we can use these two parameters as input of prediction
# Data Analysis - RIC Test Simulated Data

###### tags: `Independent Study`

:::success

- Goal:
This analysis is use the data generated in o-ran open source code.
The goal is to find the parameters used to know what can affecct UE throughput. The following is a list of data that will be analyzed :
- Data analyisis:

- 1. AD Vaildation Data
- 2. SIM E2 Data (including cell data and ue data)
- 3. AD UE Data
- 4. AD UE Anomaly Prediction Data

- Note:
- 1. **不知道AD的UE Report是在什麼環境下量測的，而且只有UE的Report，沒有Cell的資料(QP那邊的資料和AD的時間不同，不能拿來跑統計)**
- 2. **QP的Cell Report有設定RIC Test的參數**
- 3. **SIM E2 Data是假設在一輛車子移動的狀況，其他行人沒有移動的狀況下量測的**
- 4. **這裡所謂的Correlation是線性相關，也許資料之間有非線性相關，但是在Near-RT RIC少於1秒處理的狀況下，使用非線性的預測會很耗時，所以我認為如果要做thorughput的預測，必須使用linear regression**

- Conclusion:

- 1. **[AD目前的Anomaly Detection似乎不準確，和標好的資料相比，相同的數據正確的比例只有35% ](https://drive.google.com/file/d/1As1evb4_UuTmUQSg6loy7l8fh1XZeJRS/view?usp=sharing)** (Anomaly_x為validation data , Anomaly_y為經過目前AD xApp演算法後的輸出)
- 2. **無法知道這四種資料是在什麼設定下測量的，有的throughput和rsrp rsrq rssinr較有關 (>0.4)，有的較無關(~0.1)，而rsrp rsrq rssinr在不同的資料中相關係數也不同，有的比較高，有的比較低，不太好決定到底能不能是影響throguhput重大的參數**
- 3. **availPrbDl 應該也和當作throughput有關，但是在SIM E2的分析中availPrbDl 越大 throughput卻更小，不符合假設，待觀察**
- 4. **目前暫定以prb_usage, pdcpBytesDl, rsrp, rsrq, rssinr, Anomaly與UE Throughput有關**

:::
:::warning
Ref:
[Collinearity ](https://medium.com/future-vision/collinearity-what-it-means-why-its-bad-and-how-does-it-affect-other-models-94e1db984168)

:::
[TOC]

## 假設
假設下列變數與throughput呈現相關

- 1. RSRP (正相關)
- 理由: RSRP 為 UE 量測 Downlink Reference Signals (DRS) 的平均訊號強度，而DRS 為基地台在下行訊號中，於固定 RB (Resource Block) 上傳送的訊號，訊號強度越大就可以帶著更多的訊息，所以throughput也越大

- 2. RSRQ (正相關)
- 理由: RSRQ 為 從屬基地台的訊號強度與 從屬基地台的訊號強度加干擾 的比例值，如果干擾越少(RSRQ越大)，表示能傳達到的訊息也越多，所以throughput也越大

- 3. RSSINR (正相關)
- 理由: RSSINR 為 頻寬內的參考信號功率與干擾功率的比值，反映當前道路的狀況，如果道路狀況越佳，表示能傳達到的訊息也越多，所以throughput也越大

- 4. prb_usage (正相關)
- 理由: prb_usage 為 實體層定義的最小資訊的訊號單位，如果我們在便利商店有越多的櫃台可以結帳(prb_usage越多)，那throughput也當然就越大

- 5. pdcpBytesDl (正相關)
- 理由: PDCP layer 負責header壓縮與解壓縮, 資料加密/解密, 以及資料完整性保護，將重組/排序封包為原本的資料 (data) 格， 而 pdcpBytesDl 為 PDCP layer 中整個cell處理的Bytes量，如果處理的Bytes越多，資料傳輸得更快，那throughput也就越大

- 6. PrbAvailDl (正相關)
- 理由: 實體層中能夠提供的PRB數量，如果能夠提供的越多，就能傳輸更多訊息，那throughput也就越大

- 7. Anomaly (負相關)
- 理由: 在 AD xApp裡面，如果 預測結果是 Anomaly 則標示出小於threshold的參數(rsrp, rsrq, rssinr,, throughput, prb_usage)，照這樣看是負相關沒錯，不過使用的isolation forest只是區分出和正常資料不一樣的資料，所以Anomaly也可以表示高於threshold，但這邊就以xApp的程式為參考

## 探討方式 – 資料蒐集與分析工具

分析資料 & 來源:

- 1. AD Vaildation Data - https://gerrit.o-ran-sc.org/r/gitweb?p=ric-app/ad.git;a=blob;f=ad/valid.csv;h=6f09a6729b2d3c3459fe637f474800995ad69680;hb=4674f4ac8cbb4f1ca549f6acf32555d9d14c28cb
- 2. SIM E2 Data (including cell data and ue data) - https://github.com/o-ran-sc/sim-e2-interface-data
- 3. AD UE Data - https://gerrit.o-ran-sc.org/r/gitweb?p=ric-app/ad.git;a=blob;f=ad/ue.json.gz;h=0deb1bd27e939494c802181eb84c1a930c33ac2e;hb=4674f4ac8cbb4f1ca549f6acf32555d9d14c28cb
- 4. AD UE Anomaly Prediction Data - Prediction Result of AD UE Data

自變項:

- 1. RSRP

- 2. RSRQ

- 3. RSSINR

- 4. prb_usage

- 5. pdcpBytesDl

- 6. PrbAvailDl

- 7. Anomaly

依變項: UE Downlink Throughput

分析工具: 相關係數分析

## Data PreProcess Concept

The Process:

- 0. Convert to Dataframe if the data is not in Dataframe data type
- 1. Drop the data who are not useful for the prediction(UEID, Category, & Timestamp)
- 2. Extract numerical data
- 3. Drop observations having nan values
- 4. Drop observations having collinearity and have less correlation with UE throughput
- 5. Drop duplicate data

## AD Vaildation Data

### Stap 1 - Data PreProcess

import pandas as pd
import numpy as np


class PREPROCESS(object):
    r""" This PREPROCESS class takes raw data and apply prepocessing on to that.

    Parameters
    ----------
    data: pandas dataframe
        input dataset to process in pandas dataframe

    Attributes
    ----------
    data: DataFrame
        DataFrame that has processed data
    temp: list
        list of attributes to drop
    """

    def __init__(self, data):
        """
           Columns that are not useful for the prediction will be dropped(UEID, Category, & Timestamp)
        """
        self.temp = None
        self.data = data

    def variation(self):
        """ drop the constant parameters """
        if len(self.data) > 1:
            self.data = self.data.loc[:, self.data.apply(pd.Series.nunique) != 1]

    def numerical_data(self):
        """  Filters only numeric data types """
        numerics = ['int16', 'int32', 'int64', 'float16', 'float32', 'float64']
        self.data = self.data.select_dtypes(include=numerics)

    def drop_na(self):
        """ drop observations having nan values """
        self.data = self.data.dropna(axis=0)

    def process(self):
        """
          Calls the modules for the data preprocessing like dropping columns, normalization etc.,
        """
        temp_1 = ['du-id',  'measTimeStampRf', 'targetTput', 'nrCellIdentity']
              
        for col in self.data.columns:
            if 'nb' in col:
                temp_1.append(col)

        if set(temp_1).issubset(self.data.columns):
            self.temp_1 = self.data[temp_1]
            self.data = self.data.drop(temp_1, axis=1)


        
        self.numerical_data()
        self.drop_na()
        self.variation()
        #self.data.drop_duplicates()
        
        return self.data

### Step 2 - Read & Process Data

df = pd.read_csv('valid.csv')
df_ad_valid=df
df_ad_valid_process = PREPROCESS(df_ad_valid)

### Outcome 1

df_ad_valid_process.process().drop_duplicates().corr().style.background_gradient(cmap='coolwarm')


- 0. drop neighbor cell's information(rsrp rssinr rsrq) since I don't know how does the environment was set, and actually I does do the correlation with throughput but the correlation doesn't more than 0.2
- 1. throughput here is ue's throughput ( I guess is DRB.UEThpDl in RIC Test, Many commands in the oran source code mentions that we just predict downlink so far ).
- 2. we could see rsrp and rssinr has high correlation 0.98, means that they have collinearity which implies one of it is redundant, so I will remove it and remove duplicate data to calcuate the correlation again...

### Outcome 2

df_ad_valid_process.process().drop('rssinr',axis=1).drop_duplicates().corr().style.background_gradient(cmap='coolwarm')


- 1. The result shows that prb_usage and anomaly may be thoughput's important independent variables

## SIM E2 Data (including cell data and ue data)

### Step 0.1 - Explode the json & Save as csv file (SIM UE Data)

jsonToTable_iter = 8 # iterative time  (approximately)
def explode(df):
    for col in df.columns:
        if isinstance(df.iloc[0][col], list):
            df = df.explode(col)
        d = df[col].apply(pd.Series)
        df[d.columns] = d
        df = df.drop(col, axis=1)
    return df


def jsonToTable(df):
    df.index = range(len(df))
    cols = [col for col in df.columns if isinstance(df.iloc[0][col], dict) or isinstance(df.iloc[0][col], list)]
    if len(cols) == 0:
        return df
    for col in cols:
        d = explode(pd.DataFrame(df[col], columns=[col]))
        d = d.dropna(axis=1, how='all')
        df = pd.concat([df, d], axis=1)
        df = df.drop(col, axis=1).dropna()
    print(df)
    global jsonToTable_iter
    if jsonToTable_iter >= 1 :
        jsonToTable_iter = jsonToTable_iter - 1 
        return jsonToTable(df)
    else:
        return df
    
df = pd.read_json('./MeasReports-20201112-101011.json.gz', lines=True)
df = df[['ueMeasReport']] # delete the null space
df = jsonToTable(df)
df.to_csv('MeasReport_ue.csv')

- 1. I do this here is because I can't combine them as a csv file directly, so I extrace them and combine them.

### Step 0.2 - Extract Cell Data

### Step 0.3 - Explode the json & Save as csv file (SIM Cell Data)

jsonToTable_iter = 8 # iterative time  (approximately)
def explode(df):
    for col in df.columns:
        if isinstance(df.iloc[0][col], list):
            df = df.explode(col)
        d = df[col].apply(pd.Series)
        df[d.columns] = d
        df = df.drop(col, axis=1)
    return df


def jsonToTable(df):
    df.index = range(len(df))
    cols = [col for col in df.columns if isinstance(df.iloc[0][col], dict) or isinstance(df.iloc[0][col], list)]
    if len(cols) == 0:
        return df
    for col in cols:
        d = explode(pd.DataFrame(df[col], columns=[col]))
        d = d.dropna(axis=1, how='all')
        df = pd.concat([df, d], axis=1)
        df = df.drop(col, axis=1).dropna()
    print(df)
    global jsonToTable_iter
    if jsonToTable_iter >= 1 :
        jsonToTable_iter = jsonToTable_iter - 1 
        return jsonToTable(df)
    else:
        return df
    
df = jsonToTable(df)
df.to_csv('MeasReport_cell.csv')  

### Step 0.4 - Merge UE & Cell Data

df = pd.read_json('./Meas_cell.json', lines=True)

df_cell = pd.read_csv('./MeasReport_cell.csv')
df_ue =  pd.read_csv('./MeasReport_ue.csv')

merge_df= pd.merge(df_ue, df_cell, on = ['measTimeStampRf','nrCellIdentity'])
merge_df.to_csv('merge_ue_cell.csv')

### Step 1 - Data Preprocess

import pandas as pd
import numpy as np


class PREPROCESS(object):
    r""" This PREPROCESS class takes raw data and apply prepocessing on to that.

    Parameters
    ----------
    data: pandas dataframe
        input dataset to process in pandas dataframe

    Attributes
    ----------
    data: DataFrame
        DataFrame that has processed data
    temp: list
        list of attributes to drop
    """

    def __init__(self, data):
        """
           Columns that are not useful for the prediction will be dropped(UEID, Category, & Timestamp)
        """
        self.temp = None
        self.data = data

    def variation(self):
        """ drop the constant parameters """
        if len(self.data) > 1:
            self.data = self.data.loc[:, self.data.apply(pd.Series.nunique) != 1]

    def numerical_data(self):
        """  Filters only numeric data types """
        numerics = ['int16', 'int32', 'int64', 'float16', 'float32', 'float64']
        self.data = self.data.select_dtypes(include=numerics)

    def drop_na(self):
        """ drop observations having nan values """
        self.data = self.data.dropna(axis=0)

    def process(self):
        """
          Calls the modules for the data preprocessing like dropping columns, normalization etc.,
        """
        temp_1 = ['Unnamed: 0_x',
                'du-id_x', 
                'measTimeStampRf', 
                'ue-id', 
                'nrCellIdentity', 
                'targetTput']
              
        for col in self.data.columns:
            if 'nb' in col:
                temp_1.append(col)

        if set(temp_1).issubset(self.data.columns):
            self.temp_1 = self.data[temp_1]
            self.data = self.data.drop(temp_1, axis=1)
        temp_2 = ['Unnamed: 0_y','du-id_y']
  
        if set(temp_2).issubset(self.data.columns):
            self.temp_2 = self.data[temp_2]
            self.data = self.data.drop(temp_2, axis=1)
        temp_3 = ['pdcpBytesUl','availPrbUl']

        if set(temp_3).issubset(self.data.columns):
            self.temp_3 = self.data[temp_3]
            self.data = self.data.drop(temp_3, axis=1)
        temp_4 = ['rsrp.1','rssinr.1']

        if set(temp_4).issubset(self.data.columns):
            self.temp_4 = self.data[temp_4]
            self.data = self.data.drop(temp_4, axis=1)
        self.numerical_data()
        self.drop_na()
        self.variation()
    
        
        return self.data

### Step 2 - Read Merge Data and Process it

df = pd.read_csv('merge_ue_cell.csv')
merge_df = df
merge_df_processed = PREPROCESS(merge_df)

### Outcome 1

merge_df_processed.process().drop('throughput_y',axis=1).drop_duplicates().corr().style.background_gradient(cmap='coolwarm')


- 1. Maybe the pdcpBytesDl and availPrbDl can be considered to be independant variables of throughput

## AD UE Data

### Step 0 - Explode the json & Save as csv file (AD UE Data)

jsonToTable_iter = 8 # iterative time  (approximately)
def explode(df):
    for col in df.columns:
        if isinstance(df.iloc[0][col], list) and col != 'neighbourCellList':
            df = df.explode(col)
        d = df[col].apply(pd.Series)
        if col in list(range(5)):
            d.columns = d.columns + '_' + str(col)
        elif 'nbCellRfReport_' in col:
            d.columns = d.columns + '_nb_' + col[-1]
        df[d.columns] = d
        df = df.drop(col, axis=1)
    return df


def jsonToTable(df):
    df.index = range(len(df))
    cols = [col for col in df.columns if isinstance(df.iloc[0][col], dict) or isinstance(df.iloc[0][col], list)]
    if len(cols) == 0:
        return df
    for col in cols:
        d = explode(pd.DataFrame(df[col], columns=[col]))
        d = d.dropna(axis=1, how='all')
        df = pd.concat([df, d], axis=1)
        df = df.drop(col, axis=1).dropna()
    print(df)
    global jsonToTable_iter
    if jsonToTable_iter >= 1 :
        jsonToTable_iter = jsonToTable_iter - 1 
        return jsonToTable(df)
    else:
        return df
    
df = pd.read_json('./ue.json.gz', lines=True)
df = df[['ueMeasReport']] # delete the null space
df_ad = jsonToTable(df)
df_ad.to_csv('ad_ue_report.csv')

### Step 1 - Data PreProcess

import pandas as pd
import numpy as np


class PREPROCESS(object):
    r""" This PREPROCESS class takes raw data and apply prepocessing on to that.

    Parameters
    ----------
    data: pandas dataframe
        input dataset to process in pandas dataframe

    Attributes
    ----------
    data: DataFrame
        DataFrame that has processed data
    temp: list
        list of attributes to drop
    """

    def __init__(self, data):
        """
           Columns that are not useful for the prediction will be dropped(UEID, Category, & Timestamp)
        """
        self.temp = None
        self.data = data

    def variation(self):
        """ drop the constant parameters """
        if len(self.data) > 1:
            self.data = self.data.loc[:, self.data.apply(pd.Series.nunique) != 1]

    def numerical_data(self):
        """  Filters only numeric data types """
        numerics = ['int16', 'int32', 'int64', 'float16', 'float32', 'float64']
        self.data = self.data.select_dtypes(include=numerics)

    def drop_na(self):
        """ drop observations having nan values """
        self.data = self.data.dropna(axis=0)

    def process(self):
        """
          Calls the modules for the data preprocessing like dropping columns, normalization etc.,
        """
        temp_ad = ['du-id', 'measTimeStampRf', 'ue-id', 'nrCellIdentity', 'targetTput']
        
        for col in self.data.columns:
            if 'nb' in col:
                temp_ad.append(col)
        
        if set(temp_ad).issubset(self.data.columns):
            self.temp_ad = self.data[temp_ad]
            self.data = self.data.drop(temp_ad, axis=1)
            #print(self.temp_ad)
        self.numerical_data()
        self.drop_na()
        self.variation()
    
        
        return self.data

### Step 2 - Read Merge Data and Process it

df = pd.read_csv('ad_ue_report.csv')
ad_ue_report = df
ad_ue_report_df = PREPROCESS(ad_ue_report)
ad_data = ad_ue_report_df.process().drop_duplicates()

### Outcome 1

ad_data.corr().style.background_gradient(cmap='coolwarm')


- 1. throughput here is UE's throughput

- 2. rssinr has collinerity with rsrp, so I dropped it.
### Outcome 2

ad_data_1  =  ad_data.drop('rssinr',axis=1).drop_duplicates()
ad_data_1.corr().style.background_gradient(cmap='coolwarm')


## AD Anomaly Prediction Data

### Step 0 - Modify the source code of ad main.py

'''
Ken 
'''
ad_predict_df = pd.DataFrame()


def predict_anomaly(self, df):


    pred = ad_predict(df)
    df['Anomaly'] = pred
    df['Degradation'] = ''
    val = None
    '''
    Ken 
    '''
    global ad_predict_df
    ad_predict_df.append(df)
    ad_predict_df.to_csv('./ad_pred.csv')

### Stap 1 - Data PreProcess

import pandas as pd
import numpy as np


class PREPROCESS(object):
    r""" This PREPROCESS class takes raw data and apply prepocessing on to that.

    Parameters
    ----------
    data: pandas dataframe
        input dataset to process in pandas dataframe

    Attributes
    ----------
    data: DataFrame
        DataFrame that has processed data
    temp: list
        list of attributes to drop
    """

    def __init__(self, data):
        """
           Columns that are not useful for the prediction will be dropped(UEID, Category, & Timestamp)
        """
        self.temp = None
        self.data = data

    def variation(self):
        """ drop the constant parameters """
        if len(self.data) > 1:
            self.data = self.data.loc[:, self.data.apply(pd.Series.nunique) != 1]

    def numerical_data(self):
        """  Filters only numeric data types """
        numerics = ['int16', 'int32', 'int64', 'float16', 'float32', 'float64']
        self.data = self.data.select_dtypes(include=numerics)

    def drop_na(self):
        """ drop observations having nan values """
        self.data = self.data.dropna(axis=1)

    def process(self):
        """
          Calls the modules for the data preprocessing like dropping columns, normalization etc.,
        """
        temp_1 = ['Unnamed: 0','du-id',  'measTimeStampRf', 'targetTput', 'nrCellIdentity']
              
        for col in self.data.columns:
            if 'nb' in col:
                temp_1.append(col)

        if set(temp_1).issubset(self.data.columns):
            self.temp_1 = self.data[temp_1]
            self.data = self.data.drop(temp_1, axis=1)


        
        self.numerical_data()
        self.drop_na()
        self.variation()
        #self.data.drop_duplicates()
        
        return self.data

### Step 2 - Read & Process Data

ad_pred_csv = pd.read_csv('predict_outcome_df.csv')
ad_pred_df = PREPROCESS(ad_pred_csv)

### Outcome 1

ad_pred_df.process().drop_duplicates().corr().style.background_gradient(cmap='coolwarm')


- 1. rsrp and rssinr has collinearity, while rssinr has more stronger correlation with throughput, so rsrp was dropped.

### Outcome 2

ad_pred_df.process().drop('rsrp',axis=1).drop_duplicates().corr().style.background_gradient(cmap='coolwarm')


- 1. Anomaly has positive correlation with throughput is very weird.

## 結果

### AD Validation

### SIM E2

### AD UE Report

### AD UE Anomaly Prediction

## 結論

- 證實假設:

- 正相關
- RSRP : 與throughput呈現 無相關(<0.1) 或 低相關(0.1~0.4) 或 中度相關 (0.4~0.7)
- RSRQ : 與throughput呈現 無相關(<0.1) 或 低相關(0.1~0.4) 或 中度相關 (0.4~0.7)
- RSSINR: 與throughput呈現 無相關(<0.1) 或 低相關(0.1~0.4) 或 中度相關 (0.4~0.7)
- prb_usage: 與throughput呈現 中度相關 (0.4~0.7) 或 高度相關 (>0.7)
- pdcpBytesDl: 與throughput呈現 低相關(0.1~0.4)

- 負相關
- Anomaly: 與throughput呈現 中度相關 (0.4~0.7)
- 推翻假設:

- 正相關 變 負相關
- PrbAvailDl: 與throughput呈現 低相關(0.1~0.4)

- 研究結果:
- 1. RSRP RSRQ RSSINR 與 throughput關係不穩定 ?
- 2. PrbAvailDl應該要與throughput呈現正相關才對 ?
- 3. AD目前的Anomaly Prediction不準確，和標好的資料相比，相同的數據正確的比例只有35%，所以可以在AD UE Anomaly Prediction的係數表看到，Anomaly的值是和throughput呈現正相關...

暫定以 prb_usage, pdcpBytesDl, rsrp, rsrq, rssinr, Anomaly作為throughput的預測

## Apendix - QP RIC Test Configuration

{"metadata":{"simulator":
{"PRODUCT":"VIAVI RIC Test","VERSION":"v0.9-beta2-bigfile.7991-b49c9b56",
"BOOT":"2021-07-02T12:16:51.881578","CPUINFO":"4 x model name\t: unknown",
"KERNEL":"Linux version 4.4.0 #1 SMP Sun Jan 10 15:06:54 PST 2016","PROJECT":"e2-sim"}
,"configuration":{"_posted on":"2021-07-02T13:06:41.802397",
"global":{"workers":3,"dusim_gui":"","connection_points":[]},
"CTSM":{"cell_model":[{"du-id":1001,"nrCellIdentity":1,"maxPrbDl":273,"maxPrbUl":273}
,{"du-id":1002,"nrCellIdentity":2,"maxPrbDl":273,"maxPrbUl":273}],
"ue_groups":[{"ueModelConfig":{"du-id":1000,"ueModelConfigList":
[{"nrCellIdentity":0,"servingCellRfReport":{"distance":"500"},
"neighbourCellList":[{"nbCellIdentity":0,"nbCellRfReport.distance":"13300"},
{"nbCellIdentity":1,"nbCellRfReport.distance":"500"}]}]}},
{"ueModelConfig":{"du-id":1000,"ueModelConfigList"
:[{"nrCellIdentity":0,"servingCellRfReport"
:{"rsrq":"50,60,70","rsrp":"50,60,70","rssinr":"50,60,70"},
"neighbourCellList":[{"nbCellIdentity":1,
"nbCellRfReport":{"rsrq":"50,60,70","rsrp":"50,60,70","rssinr":"50,60,70"}}
,{"nbCellIdentity":2,"nbCellRfReport":
{"rsrq":"50,60,70","rsrp":"50,60,70","rssinr":"50,60,70"}}]}]}}]},
"Simulation":{"cells":{"number":13,"neighbours":5,"antenna_height":10,
"distribution":{"type":"diagram","distance":1000,"diagram":"           
     c6\n              a1~~~~\n          c7  ~~~~~~  c5\n          
         ~~~~~a1\n    c8          c1\n            a2-------·       
             \n          c2¦        ¦c4\n    b1‾‾‾‾‾‾‾‾|      ¦    
                   a3\n  c9|T.Station|c3    ¦      c13  a3\n    |_________b1   
                     ¦\n        c10¦         ¦c12\n            ·--------a2\n     
                                c11","Street width":20},
                                "layers":[{"name":"B2","frequency":1900,
                                "power":3,"maxPrbDl":120,"maxPrbUl":120},
                                {"name":"B13","frequency":700,"power":-1,
                                "maxPrbDl":91,"maxPrbUl":91},
                                {"name":"N77","frequency":3600,"power":20,
                                "maxPrbDl":273,"maxPrbUl":273}]},
                                "type2_ues":[{"global-id":"Pedestrian-{n}",
                                "Description":"Slow UEs using 250 Mbps, 
                                most around cell 1","targetTput":0.25,
                                "handover":"RSRP (threshold=5)",
                                "mobility":{"type":"short range","speed":0.5,
                                "roundtrip":"return",
                                "distribution":[{"ues":10,"locations":
                                "c1/B2,c1/B2,c1/B2,c2/B2,c3/B2,c4/B2,c5/B2,
                                c6/B2,c7/B13"}]}},{"global-id":"Car-{n}",
                                "Description":"A car driving North from cell 1",
                                "targetTput":0.75,"handover":"RSRP (threshold=5)",
                                "mobility":{"type":"Manhattan","speed":20,
                                "roundtrip":"teleport","distribution":
                                [{"ues":1,"locations":"c1/B13"}],
                                "path":"NNWSEES..."}},{"global-id":
                                "Train passenger {n}","Description":
                                "Passengers trains arriving from the East, 
                                then scattering from the Train Station","targetTput":0.30000000000000004,"handover":"RSRP (threshold=5)",
                                "mobility":{"type":"Manhattan","speed":10,"roundtrip":"teleport","distribution":[{"ues":3,"locations":"a3"}],"path":"WW..."}}
                                ,{"global-id":"Waiting passenger {n}","Description":"Passengers waiting for their train in the Train Station","targetTput":0.1,"
                                handover":"RSRP (threshold=5)","mobility":{"type":"building","distribution":[{"ues":10,"locations":"b1"}]}}],"buildings"
                                :{"rf_degradation":{"rsrp_rsqr_loss":20,"rssinr_loss":10},"types":[{"height":20,"floors":4},{"height":50,"floors":10}]}},
                                "System":{"report_ms":10,"generate_ms":10,"batch_mode":true,"batch_length":2000,"tcp_hostport":"127.0.0.1:3001",
                                "RIC_address":"127.0.0.1:36422","Send_simulation_data":false,"PLMN_id":"001f01","gNB_id":"12345c"},
                                "Models":{"cell_models":[{"du-id":1001,"nrCellIdentity":"c1/B2",
                                "nrCellIdentity5G":4096,"maxPrbDl":120,"maxPrbUl":120,"frequency":1900,
                                "power":3,"xyz":"0,0,10"},{"du-id":1001,"nrCellIdentity":"c1/B13",
                                "nrCellIdentity5G":4097,"maxPrbDl":91,"maxPrbUl":91,"frequency":700,
                                "power":-1,"xyz":"0,0,10"},{"du-id":1001,"nrCellIdentity":"c1/N77",
                                "nrCellIdentity5G":4098,"maxPrbDl":273,"maxPrbUl":273,"frequency":3600,
                                "power":20,"xyz":"0,0,10"},{"du-id":1002,"nrCellIdentity":"c2/B2","nrCellIdent
                                ity5G":8192,"maxPrbDl":120,"maxPrbUl":120,"frequency":1900,"power":3,"xyz":"-832,-55
                                ,10"},{"du-id":1002,"nrCellIdentity":"c2/B13","nrCellIdentity5G":8193,"maxPrbDl":91,"m
                                axPrbUl":91,"frequency":700,"power":-1,"xyz":"-832,-555,10"},{"du-id":1002,"nrCellIdentity":
                                "c2/N77","nrCellIdentity5G":8194,"maxPrbDl":273,"maxPrbUl":273,"frequency":3600,"power":20,"xyz":"-832,-555,10"},
                                {"du-id":1003,"nrCellIdentity":"c3/B2","nrCellIdentity5G":12288,"maxPrbDl":120,"maxPrbUl":120,"frequency":1900
                                ,"power":3,"xyz":"-139,-1109,10"},{"du-id":1003,"nrCellIdentity":"c3/B13","nrCellIdentity5G":12289,"maxPrbDl":91,
                                "maxPrbUl":91,"frequency":700,"power":-1,"xyz":"-139,-1109,10"},{"du-id":1003,"nrCellIdentity":"c3/N77",
                                "nrCellIdentity5G":12290,"maxPrbDl":273,"maxPrbUl":273,"frequency":3600,"power":20,"xyz":"-139,-1109,10"},
                                {"du-id":1004,"nrCellIdentity":"c4/B2","nrCellIdentity5G":16384,"maxPrbDl":120,"maxPrbUl":120,"frequency":1900,"power":3,"
                                xyz":"832,-555,10"},{"du-id":1004,"nrCellIdentity":"c4/B13","nrCellIdentity5G":16385,"maxPrbDl":91,"maxPrbUl":91,
                                "frequency":700,"power":-1,"xyz":"832,-555,10"},{"du-id":1004,"nrCellIdentity":"c4/N77","nrCellIdentity5G":16386,"maxPrbDl":273
                                ,"maxPrbUl":273,"frequency":3600,"power":20,"xyz":"832,-555,10"},{"du-id":1005,"nrCellIdentity":"c5/B2",
                                "nrCellIdentity5G":20480,"maxPrbDl":120,"maxPrbUl":120,"frequency":1900,"power":3,"xyz":"832,555,10"},
                                {"du-id":1005,"nrCellIdentity":"c5/B13","nrCellIdentity5G":20481,"maxPrbDl":91,"maxPrbUl":91,
                                "frequency":700,"power":-1,"xyz":"832,555,10"},{"du-id":1005,"nrCellIdentity":"c5/N77","nrCellIdentity5G":20482,"
                                maxPrbDl":273,"maxPrbUl":273,"frequency":3600,"power":20,"xyz":"832,555,10"},{"du-id":1006,"nrCellIdentity":"c6/B2","nrCellIdentity5G":24576,"maxPrbDl":120,"maxPrbUl":120,
                                "frequency":1900,"power":3,"xyz":"0,1109,10"},{"du-id":1006,"nrCellIdentity":"c6/B13","nrCellIdentity5G":24577,"maxPrbDl":91,"maxPrbUl":91,"frequency":700,"power":-1,"xyz":"0,1109,10"}
                                ,{"du-id":1006,"nrCellIdentity":"c6/N77","nrCellIdentity5G":24578,"maxPrbDl":273,"maxPrbUl":273,"frequency":3600,"power":20,"xyz":"0,1109,10"},{"du-id":1007,"nrCellIdentity":"c7/B2",
                                
                                "nrCellIdentity5G":28672,"maxPrbDl":120,"maxPrbUl":120,"frequency":1900,"power":3,"xyz":"-832,555,10"},{"du-id":1007,"nrCellIdentity":"c7/B13","nrCellIdentity5G":28673,"maxPrbDl":91,
                                "maxPrbUl":91,"frequency":700,"power":-1,"xyz":"-832,555,10"},{"du-id":1007,"nrCellIdentity":"c7/N77","nrCellIdentity5G":28674,"maxPrbDl":273,"maxPrbUl":273,"frequency":3600,"power":20,
                                "xyz":"-832,555,10"},{"du-id":1008,"nrCellIdentity":"c8/B2","nrCellIdentity5G":32768,"maxPrbDl":120,"maxPrbUl":120,"frequency":1900,"power":3,"xyz":"-1664,0,10"},{"du-id":1008,
                                "nrCellIdentity":"c8/B13","nrCellIdentity5G":32769,"maxPrbDl":91,"maxPrbUl":91,"frequency":700,"power":-1,"xyz":"-1664,0,10"},{"du-id":1008,"nrCellIdentity":"c8/N77"
                                ,"nrCellIdentity5G":32770,"maxPrbDl":273,"maxPrbUl":273,"frequency":3600,"power":20,"xyz":"-1664,0,10"},{"du-id":1009,"nrCellIdentity":"c9/B2","nrCellIdentity5G":36864,"maxPrbDl":120,
                                "maxPrbUl":120,"frequency":1900,"power":3,"xyz":"-1941,-1109,10"},{"du-id":1009,"nrCellIdentity":"c9/B13","nrCellIdentity5G":36865,"maxPrbDl":91,"maxPrbUl":91,"frequency":700,"power":-1,
                                "xyz":"-1941,-1109,10"},{"du-id":1009,"nrCellIdentity":"c9/N77","nrCellIdentity5G":36866,"maxPrbDl":273,"maxPrbUl":273,"frequency":3600,"power":20,"xyz":"-1941,-1109,10"},{"du-id":1010,
                                "nrCellIdentity":"c10/B2","nrCellIdentity5G":40960,"maxPrbDl":120,"maxPrbUl":120,"frequency":1900,"power":3,"xyz":"-1109,-1664,10"},{"du-id":1010,"nrCellIdentity":"c10/B13",
                                "nrCellIdentity5G":40961,"maxPrbDl":91,"maxPrbUl":91,"frequency":700,"power":-1,"xyz":"-1109,-1664,10"},{"du-id":1010,"nrCellIdentity":"c10/N77","nrCellIdentity5G":40962,"maxPrbDl":273
                                "maxPrbUl":273,"frequency":3600,"power":20,"xyz":"-1109,-1664,10"},{"du-id":1011,"nrCellIdentity":"c11/B2","nrCellIdentity5G":45056,"maxPrbDl":120,"maxPrbUl":120,"frequency":1900,
                                "power":3,"xyz":"0,-2219,10"},{"du-id":1011,"nrCellIdentity":"c11/B13","nrCellIdentity5G":45057,"maxPrbDl":91,"maxPrbUl":91,"frequency":700,"power":-1,"xyz":"0,-2219,10"},
                                {"du-id":1011,"nrCellIdentity":"c11/N77","nrCellIdentity5G":45058,"maxPrbDl":273,"maxPrbUl":273,"frequency":3600,"power":20,"xyz":"0,-2219,10"},{"du-id":1012,"nrCellIdentity":"c12/B2",
                                "nrCellIdentity5G":49152,"maxPrbDl":120,"maxPrbUl":120,"frequency":1900,"power":3,"xyz":"832,-1664,10"},{"du-id":1012,"nrCellIdentity":"c12/B13","nrCellIdentity5G":49153,
                                "maxPrbDl":91,"maxPrbUl":91,"frequency":700,"power":-1,"xyz":"832,-1664,10"},{"du-id":1012,"nrCellIdentity":"c12/N77","nrCellIdentity5G":49154,"maxPrbDl":273,"maxPrbUl":273,
                                "frequency":3600,"power":20,"xyz":"832,-1664,10"},{"du-id":1013,"nrCellIdentity":"c13/B2","nrCellIdentity5G":53248,"maxPrbDl":120,"maxPrbUl":120,"frequency":1900,"power":3,
                                "xyz":"1664,-1109,10"},{"du-id":1013,"nrCellIdentity":"c13/B13","nrCellIdentity5G":53249,"maxPrbDl":91,"maxPrbUl":91,"frequency":700,"power":-1,"xyz":"1664,-1109,10"}
                                {"du-id":1013,"nrCellIdentity":"c13/N77","nrCellIdentity5G":53250,"maxPrbDl":273,"maxPrbUl":273,"frequency":3600,"power":20,"xyz":"1664,-1109,10"}],"type2_groups":[
                                    {"du-id":1004,"ueModelConfigList":[{"ue-id":"Pedestrian-1","nrCellIdentity":"c4/B2","targetTput":0.25,"xyz":"731,-72","waypoints":"731,-72;731,1038","path":"",
                                    "speed":0.5,"roundtrip":"return","handover":"RSRP (threshold=5)"},{"ue-id":"Pedestrian-3","nrCellIdentity":"c4/B2","targetTput":0.25,"xyz":"1470,315","waypoints"
                                    "1470,315;1470,-795","path":"","speed":0.5,"roundtrip":"return","handover":"RSRP (threshold=5)"},{"ue-id":"Pedestrian-7","nrCellIdentity":"c4/B2",
                                    "targetTput":0.25,"xyz":"1593,-31","waypoints":"1593,-31;1593,1079","path":"","speed":0.5,"roundtrip":"return","handover":"RSRP (threshold=5)"},{
                                        "
                                    ue-id":"Pedestrian-10","nrCellIdentity":"c4/B2","targetTput":0.25,"xyz":"1775,325","waypoints":"1775,325;1775,-785","path":"","speed":0.5,"roundtr

                                    ip":"return","handover":"RSRP (threshold=5)"}]},{"du-id":1005,"ueModelConfigList":[{"ue-id":"Pedestrian-2","nrCellIdentity":"c5/B2","targetTput":0.25,"xyz":"120
                                    8,284","waypoints":"1208,284;1208,-826","path":"","speed":0.5,"roundtrip":"return","handover":"RSRP (threshold=5)"}]},
{"du-id":1003,"ueModelConfigList":[{"ue-id":"Pedestrian-4","nrCellIdentity":"c3/B2","targetTput":0.25,"xyz":"-1019,-783","waypoints":"-1019,-783;-1296,-1892",
"path":"","speed":0.5,"roundtrip":"return","handover":"RSRP (threshold=5)"},{"ue-id":"Waiting passenger 4","nrCellIdentity":"c3/B2","targetTput":0.1,"xyz":"-435,-857","waypoints":"","
path":"","speed":0,"roundtrip":"return","handover":"RSRP (threshold=5)"}]},{"du-id":1002,"ueModelConfigList":[{"ue-id":"Pedestrian-5","nrCellIdentity":"c2/B2","targetTput":0.25,"xyz":
"-748,-481","waypoints":"-748,-481;-55,-1035","path":"","speed":0.5,"roundtrip":"return","handover":"RSRP (threshold=5)"},{"ue-id":"Pedestrian-8","nrCellIdentity":"c2/B2","targetTput"
:0.25,"xyz":"-503,-607","waypoints":"-503,-607;190,-1161","path":"","speed":0.5,"roundtrip":"return","handover":"RSRP (threshold=5)"},{"ue-id":"Pedestrian-9","nrCellIdentity":"c2/B2",
"targetTput":0.25,"xyz":"111,-404","waypoints":"111,-404;943,-959","path":"","speed":0.5,"roundtrip":"return","handover":"RSRP (threshold=5)"},{"ue-id":"Waiting passenger 2","nrCellIde
ntity":"c2/B2","targetTput":0.1,"xyz":"-726,-1130,5","waypoints":"","path":"","speed":0,"roundtrip":"return","handover":"RSRP (threshold=5)"},{"ue-id":"Waiting passenger 3","nrCellIden
tity":"c2/N77","targetTput":0.1,"xyz":"-691,-973,5","waypoints":"","path":"","speed":0,"roundtrip":"return","handover":"RSRP (threshold=5)"},{"ue-id":"Waiting passenger 5","nrCellIdent
ity":"c2/B2","targetTput":0.1,"xyz":"-899,-942","waypoints":"","path":"","speed":0,"roundtrip":"return","handover":"RSRP (threshold=5)"},{"ue-id":"Waiting passenger 7","nrCellIdentity":"
c2/N77","targetTput":0.1,"xyz":"-973,-1060,15","waypoints":"","path":"","speed":0,"roundtrip":"return","handover":"RSRP (threshold=5)"},{"ue-id":"Waiting passenger 8","nrCellIdentity":"c2/B2","targetTput":0.1,
"

xyz":"-1093,-886,15","waypoints":"","path":"","speed":0,"roundtrip":"return","handover":"RSRP (threshold=5)"}]},{"du-id":1007,"ueModelConfigList":[{"ue-id":"Pedestrian-6","nrCellIdentity":"c7/B13","targetTput":
0.25,"xyz":"-393,-382","waypoints":"-393,-382;439,173","path":"","speed":0.5,"roundtrip":"return","handover":"RSRP (threshold=5)"}]},{"du-id":1001,"ueModelConfigList":[{"ue-id":"Car-1","nrCellIdentity":"c1/B13"
,"targetTput":0.75,"xyz":"-567,827","waypoints":"","path":"NNWSEES...","speed":20,"roundtrip":"teleport","handover":"RSRP (threshold=5)"}]},{"du-id":1013,"ueModelConfigList":[{"ue-id":"Train passenger 1",
"nrCellIdentity":"c13/B13","targetTput":0.30000000000000004,"xyz":"2330,-977","waypoints":"","path":"WW...","speed":10,"roundtrip":"teleport","handover":"RSRP (threshold=5)"},{"ue-id":"Train passenger 2",
nrCellIdentity":"c13/B2","targetTput":0.30000000000000004,"xyz":"2347,-1103","waypoints":"","path":"WW...","speed":10,"roundtrip":"teleport","handover":"RSRP (threshold=5)"},{"ue-id":"Train passenger 3",
"nrCellIdentity":"c13/N77","targetTput":0.30000000000000004,"xyz":"2307,-898","waypoints":"","path":"WW...","speed":10,"roundtrip":"teleport","handover":"RSRP (threshold=5)"}]},{"du-id":1010,
"ueModelConfigList":[{"ue-id":"Waiting passenger 1","nrCellIdentity":"c10/N77","targetTput":0.1,"xyz":"-851,-1232,15","waypoints":"","path":"","speed":0,"roundtrip":"return","handover":"RSRP (threshold=5)"}
{"ue-id":"Waiting passenger 6","nrCellIdentity":"c10/B2","targetTput":0.1,"xyz":"-1095,-1160,5","waypoints":"","path":"","speed":0,"roundtrip":"return","handover":"RSRP (threshold=5)"}
,{"ue-id":"Waiting passenger 9","nrCellIdentity":"c10/N77","targetTput":0.1,"xyz":"-1353,-1312,10","waypoints":"","path":"","speed":0,"roundtrip":"return","handover":"RSRP (threshold=5)"}

,{"ue-id":"Waiting passenger 10","nrCellIdentity":"c10/N77","targetTput":0.1,"xyz":"-1372,-1359","waypoints":"","path":"","speed":0,"roundtrip":"return","handover":"RSRP (threshold=5)"}]}],
"areas":[{"name":"a1","points":"-277,277;416,832"},{"name":"a2","points":"-555,-1941;693,-277"},{"name":"a3","points":"2219,-1109;2357,-832"},{"name":"b1","points":"-1664,-1387;-277,-832,20"},
{"name":"b1#f1","points":"-1664,-1387;-277,-832,5"},{"name":"b1#f2","points":"-1664,-1387,5;-277,-832,10"},{"name":"b1#f3","points":"-1664,-1387,10;-277,-832,15"},{"name":"b1#f4",
"points":"-1664,-1387,15;-277,-832,20"}],"meta":{"last_seed":"0x7db2"}},"Reports":{"logging":"info","full_reporting":true,"include_internal":true,"api_issues":"fix","counts_max_seconds":3},
"Anomalies":[{"anomaly_name":"Intermittent Throughput issue in cell 1","degradation":{"type":"throughput","value":80},"delay":10,"duration":3,"affected_areas":"","affected_cells":"",
"affected_ues":"Car*, Pedes*"},{"anomaly_name":"Radio problem in a1 area (except for cell 7)","degradation":{"type":"radio","rsrp":20,"rsrq":20,"rssinr":20},"delay":8,"duration":2,
"affected_areas":"","affected_cells":"","affected_ues":"Train*"},{"anomaly_name":"Intermittent issues for cars in both areas","degradation":{"type":"radio","rsrp":20,"rsrq":20,"rssinr":20}
,"delay":10,"duration":3,"affected_areas":"","affected_cells":"","affected_ues":"Waiting*"}],"E2 Load":{"total_CUs":1,"DUs_per_CU":3,"Cells_per_DU":15,"UEs_per_Cell":32,"UE_groups"
:[{"name":"High","period":10,"ues":10},{"name":"Medium","period":60000,"ues":30},{"name":"Low","period":900000,"ues":30},{"name":"ReportEveryHour","period":3600000,"ues":10}]}},
"numbers":{"UEs":416,"cells":13,"CUs":1,"DUs":3},"tags":["radio anomaly","throughput anomaly","buildings","degradation"],"Throughput Anomaly":[{"anomaly_name":"Intermittent Throughput issue in cell 1",
"degradation":{"type":"throughput","value":80},"delay":10,"duration":3,"affected_areas":"","affected_cells":"","affected_ues":"Car*, Pedes*"}],"Radio Anomaly":[{"anomaly_name":"Radio problem in a1 area 
(except for cell 7)","degradation":{"type":"radio","rsrp":20,"rsrq":20,"rssinr":20},"delay":8,"duration":2,"affected_areas":"","affected_cells":"","affected_ues":"Train*"},{"anomaly_name":"Intermittent 
issues for cars in both areas","degradation":{"type":"radio","rsrp":20,"rsrq":20,"rssinr":20},"delay":10,"duration":3,"affected_areas":"","affected_cells":"","affected_ues":"Waiting*"}]}}


'''