Merge pull request #6 from Cortze/random-peering

Documentation Update

Author: cortze

.gitignore

@@ -8,12 +8,12 @@ src/bin
 
 # Ignore the example folder (personal to each user)
 examples/*
+# Ignore the example folder (personal to each user)
+general-results/*
 
-# Ignore System Files
+# Ignore System Files (from MAC)
 .DS_Store
 ._.DS_Store
 
 # ignore the venv
 src/analyzer/venv
-
-src/analyzer/dashboard/static/plots

README.md

@@ -9,17 +9,21 @@ To use the tool the following tools needs to be installed on the machine:
 
 ### Usage
 The crawler can be easily executed from the `armiarma.sh` file (make sure that is an executable file). 
-The executable `armiarma.sh` launching shell script supports three different commands that correspond to the main functionalities of the tool.
+The executable `armiarma.sh` launching shell script supports five different commands that correspond to the main functionalities of the tool.
 Those commands are:
 
 - `./armiarma.sh -h` to display the help text of the tool. 
 - `./armiarma.sh -c [network] [project-name]` to launch the armiarma crawler on the given network. The given project-name is the name of the folder where the gathered metrics and where all the related information will be saved. The folder will be placed on the `./examples` folder. 
 - `./armiarma.sh -p [project-name]` to launch the armiarma analyzer over the metrics from the given project. The generted plots of the analysis will be saved on the `./examples/[project-name]/plots` folder.
+- `./armiarma.sh -f [network] [project-name] [time]` to launch the armiarma crawler on the given network for the given `time` in minutes. The given project-name is the name of the folder where the gathered metrics and where all the related information will be saved. The folder will be placed on the `./examples` folder. After the crawling for the given time, the analyzer will be launched automatically. The summary results will be saved on the `./examples/[project-name]/plots` folder.
+- `./armiarma.sh -o` to launch the armiarma general overview analyzer over the metrics from the projects' folder. The generted results of the analysis will be saved on the `./general-resutls` folder.
 
 ```
     ./armiarma.sh --> -h
                   '-> -c [network] [project-name]
                   '-> -p [project-name]
+                  '-> -f [network] [project-name] [time](minutes)
+                  '-> -o 
 ```
 
 The tool will get the Go packages and compile Rumor as generate the virtual env and install the python3 dependencies for the user.  

armiarma.sh

@@ -30,13 +30,19 @@ Help()
     echo "                  *Parameters for -p [name]"
     echo "          -f      Run a time specified test, performing the analysis"
     echo "                  of the obtained"
+    echo "                  *Parameters for -f [network] [project-name] [time](minutes)"
+    echo "          -o      Run the general analysis over the entire projects' folder."
+    echo ""
     echo "      Parameters:"
     echo "          [network]       The ETH2 network where the crawler will be running"
     echo "                          Currently supported networks:"
     echo "                              -> mainnet" 
     echo "          [project-name]  Specify the name of the folder where the metrics" 
     echo "                          and plots will be stored."
     echo "                          Find them on 'armiarma/examples/[project-name]'"
+    echo "          [time]          Specific time in minutes to run the crawler, performing afterwards"
+    echo "                          the general analysis."
+    echo ""
     echo ""
     echo "  BSC-ETH2 TEAM"
 }
@@ -142,7 +148,7 @@ LaunchCrawler(){
             # Move to the example folder
             cd "./examples/${folderName}"
 
-            # Append the bash env variables to the temp file
+            # Append the bash env variables to the temp fhiustile
             echo "metricsFolder=\"${metricsFolder}\"" >> config.sh
             echo "armiarmaPath=\"${folderPath}\"" >> config.sh
 
@@ -211,7 +217,7 @@ LaunchAnalyzer(){
         echo "venv already created"
     else
         echo "Generating the virtual env"
-        python -m virtualenv "$VENV"
+        python3 -m virtualenv "$VENV"
     fi
     echo ""
     # Source the virtual env 
@@ -230,7 +236,7 @@ LaunchAnalyzer(){
     # -- END TEMP --
 
     echo "Checking if Python dependencies are installed..." 
-    pip install -r ./src/analyzer/requirements.txt
+    pip3 install -r ./src/analyzer/requirements.txt
     echo ""
 
     # Set the Paths for the gossip-metrics.json peerstore.json and output
@@ -249,20 +255,53 @@ LaunchAnalyzer(){
     # Run the Analyzer
     echo "  Launching analyzer"
     echo ""
-    python ./src/analyzer/armiarma-analyzer.py "$csv" "$peerstore" "$extrametrics" "$plots"
+    python3 ./src/analyzer/armiarma-analyzer.py "$csv" "$peerstore" "$extrametrics" "$plots"
 
     # Deactivate the VENV
     deactivate
 
 }
 
+# Launch the General Overview of all the projects in the examples folder
+LaunchGeneralResults(){
+    if [[ -d "./general-results/plots" ]]; then
+        echo ""
+    else
+        mkdir "./general-results/plots"
+    fi
+
+    # Source the virtual env 
+    source "${VENV}/bin/activate"  
+
+    # Check if the virtual env is created
+    venvPath="${PWD}/src/analyzer/venv"
+    if [[ "$VIRTUAL_ENV" = "$venvPath" ]]
+    then
+        echo "VENV successfuly sourced"
+    else
+        echo "ERROR. VENV was unable to source" >&2
+    fi
+    echo ""
+
+    # Run the Analyzer
+    echo "  Launching General Overview Analyzer"
+    echo ""
+    python3 ./src/analyzer/total-overview-analysis.py ./examples ./general-results
+    echo "results available in \$ARMIARMA/results"
+    echo ""
+    # Deactivate the VENV
+    deactivate
+}
+
+
 # -------- END OF FUNCTION DEFINITION ---------
 
 # Execution Secuence
 
 # 0. Get the options
 go version
 
+# Generate the examples folder
 if [[ -d ./examples ]]; then
     echo ""
     echo "  ----------- ARMIARMA $version -----------"
@@ -274,6 +313,15 @@ else
     echo ""
     mkdir ./examples  
 fi 
+# Generate the general-results folder
+if [[ -d ./general-results ]]; then
+    echo ""
+else
+    echo ""
+    echo "Generating ./general-results folder"
+    echo ""
+    mkdir ./general-results  
+fi
 
 # Check if any argument was given. 
 # If not, print Help and exit
@@ -283,7 +331,7 @@ if [[ -z "$1" ]]; then
     exit 1
 fi
 
-while getopts ":hcpfdts" option; do
+while getopts ":hcpfdo" option; do
     case $option in
         h)  # display Help
             Help
@@ -329,10 +377,18 @@ while getopts ":hcpfdts" option; do
 
             echo "Calling the Analyzer"
             LaunchAnalyzer "$folderName" "$folderPath"
-
+            
             echo "Exit Crawler execution"
             exit;;
 
+        o)  # Generate the general overview of the previously generated projects
+
+            LaunchGeneralResults
+
+            echo "Overview Analyzer Finished!"
+            echo ""
+            exit;;
+
         d)  # Option to run Armiarma Crawler for a specific given time on the iexec decentralice platform
             # processing the results with the Analyzer and returning the metrics and plots in a zip file
             # TODO: encrypt the results with a certain security measure, so noone can access to them

src/analyzer/armiarma-analyzer.py

@@ -752,7 +752,7 @@ def main():
 
         ## -- website code --
         print("\n")
-        print("Results from crawler run on [month] running for [crawling time].<br>Total amount of peers on the peerstore:", peerstoreLen,".<br>Number of clients with the TPC port at 13000 (Prysm?):", cnt13000,".<br>Percentage of 'Prysm' peers from the peerstore (based on the TCP port):", round((cnt13000*100)/peerstoreLen,2),"%.<br>We manage to connect with", succeed,"peers from the peerstore.<br>This would be the distribution.")
+        print("Results from crawler run on [month] running for [crawling time].\n<br>Total amount of peers on the peerstore:", peerstoreLen,".\n<br>Number of clients with the TPC port at 13000 (Prysm?):", cnt13000,".\n<br>Percentage of 'Prysm' peers from the peerstore (based on the TCP port):", round((cnt13000*100)/peerstoreLen,2),"%.\n<br>We manage to connect with", succeed,"peers from the peerstore.\n<br>This would be the distribution.")
         print("\n")
 
 
@@ -818,15 +818,15 @@ def main():
             'textSize': textSize+2,                                                         
             'yLowLimit': None,                                                             
             'yUpperLimit': None,                                                        
-            'title': "Distribution of the detected errors",                             
-            'xlabel': 'Number of peers',                                   
+            'title': "Distribution of the Experienced Connection Errors",                             
+            'xlabel': 'Peers With Same Connection Error',                                   
             'ylabel': None,                                                
             'yticks': None, 
             'vGrids': True,                                                           
             'titleSize': titleSize+2,                                                        
             'labelSize': labelSize+2,                                                        
             'legendPosition': 1,
-            'legendTitle': 'Experienced errors',                                                   
+            'legendTitle': 'Recorded errors',                                                   
             'legendSize': labelSize-4,                                                       
             'xticksSize': ticksSize,                                                       
             'yticksSize': ticksSize+2,                                                            

src/analyzer/total-overview-analysis.py

@@ -0,0 +1,142 @@
+# This file compiles the code to analyze the gathered metrics from each of the 
+# folders in the $ARMIARMAPATH/examples folder, providing a global overview of the 
+# Peer distributions and client types over the dayly analysis
+
+import os, sys
+import json
+import pandas as pd
+import matplotlib.pyplot as plt
+import numpy as np
+
+
+def mainexecution():
+    projectsFolder = sys.argv[1]
+    outFolder = sys.argv[2]
+
+    outJson = outFolder + '/' + 'client-distribution.json'
+    # So far, the generated panda will just contain the client count
+    p = {'date': [], 'Lighthouse': [], 'Teku': [], 'Nimbus': [], 'Prysm': [], 'Lodestar': [], 'Unknown': []}
+    # Concatenation of the Json values
+    j = []
+    print(projectsFolder)
+    for root, dirs, _ in os.walk(projectsFolder):
+        print(dirs)
+        dirs.sort()
+        # We just need to access the folders inside examples
+        for d in dirs:
+            fold = os.path.join(root, d)
+            f = fold + '/metrics/custom-metrics.json'
+            if os.path.exists(f):
+                # Load the readed values from the json into the panda
+                poblatePandaCustomMetrics(p, j, f)
+        break
+
+    # Export the Concatenated json to the given folder
+    with open(outJson, 'w', encoding='utf-8') as f:
+        json.dump(j, f, ensure_ascii=False, indent=4)
+
+    # After filling the dict with all the info from the JSONs, generate the panda
+    df = pd.DataFrame (p, columns = ['date', 'Lighthouse', 'Teku', 'Nimbus', 'Prysm', 'Lodestar', 'Unknown'])
+    df = df.sort_values(by="date")
+    print(df)
+
+    outputFigsFolder = outFolder + '/' + 'plots'
+    clientList = ['Lighthouse', 'Teku', 'Nimbus', 'Prysm', 'Lodestar', 'Unknown']
+    figSize = (10,6)
+    titleSize = 24
+    textSize = 20
+    labelSize = 20
+    # Plot the images
+    plotStackedChart(df, opts={                                   
+            'figSize': figSize,                                                      
+            'figTitle': 'Client-Distribution.png',  
+            'figtitle': 'client-distribution.png',                                 
+            'outputPath': outputFigsFolder,                                        
+            'align': 'center',
+            'grid': 'y',    
+            'textSize': textSize,                                                  
+            'title': "Evolution of the projected client distribution",           
+            'ylabel': 'Client concentration (%)',                                                         
+            'xlabel': None, 
+            'yticks': np.arange(0, 110, 10),
+            'legendLabels': clientList,                    
+            'titleSize': titleSize,                                                        
+            'labelSize': labelSize,                                                        
+            'lengendPosition': 1,                                                   
+            'legendSize': labelSize,                                                     
+            'tickRotation': 90,                                                     
+            'show': True})
+
+def poblatePandaCustomMetrics(p, j, jsonFile):
+    # Read the Json
+    jsf = open(jsonFile)
+    jsonValues = json.load(jsf)
+    jsf.close()
+    
+    # Add readed Json to the previous ones (concatenate them) 
+    j.append(jsonValues)
+
+    # Compose the date of the crawling day
+    cday = str(jsonValues['StartTime']['Year']) + '-' + str(jsonValues['StartTime']['Month']) + '-' + str(jsonValues['StartTime']['Day'])
+    p['date'].append(cday)
+    ps = jsonValues['PeerStore']['Total']
+    # Get percentages for each of the clients
+    crwlig = jsonValues['PeerStore']['ConnectionSucceed']['Lighthouse']['Total']
+    crwtek = jsonValues['PeerStore']['ConnectionSucceed']['Teku']['Total']
+    crwnim = jsonValues['PeerStore']['ConnectionSucceed']['Nimbus']['Total']
+    crwlod = jsonValues['PeerStore']['ConnectionSucceed']['Lodestar']['Total']
+    crwunk = jsonValues['PeerStore']['ConnectionSucceed']['Unknown']['Total']
+
+    prysPort = jsonValues['PeerStore']['Port13000']
+    res = ps - prysPort
+
+    noPrysm = crwlig + crwtek + crwnim + crwlod + crwunk
+    
+    estimlig = (res*crwlig)/noPrysm
+    estimtek = (res*crwtek)/noPrysm
+    estimnim = (res*crwnim)/noPrysm
+    estimlod = (res*crwlod)/noPrysm
+    estimunk = (res*crwunk)/noPrysm
+
+    lig = round((estimlig*100)/ps, 2)
+    tek = round((estimtek*100)/ps, 2)
+    nim = round((estimnim*100)/ps, 2)
+    pry = round((prysPort*100)/ps, 2)
+    lod = round((estimlod*100)/ps, 2)
+    unk = round((estimunk*100)/ps, 2)
+
+    p['Lighthouse'].append(lig)
+    p['Teku'].append(tek)
+    p['Nimbus'].append(nim)
+    p['Prysm'].append(pry)
+    p['Lodestar'].append(lod)
+    p['Unknown'].append(unk)
+
+
+def plotStackedChart(p, opts):
+    outputFile = str(opts['outputPath']) + '/' + opts['figTitle']    
+
+    ax = p.plot.area(figsize = opts['figSize'], x='date', stacked=True)
+
+    # labels
+    if opts['ylabel'] is not None:     
+        plt.ylabel(opts['ylabel'], fontsize=opts['labelSize'])
+    if opts['xlabel'] is not None:
+        plt.xlabel(opts['xlabel'], fontsize=opts['labelSize'])
+    
+    handles,legends = ax.get_legend_handles_labels()
+    ax.legend(handles=handles, labels=opts['legendLabels'], loc='upper center', bbox_to_anchor=(0.5, -0.05), fancybox=True, shadow=True, ncol=6)
+
+    if opts['grid'] != None:
+        ax.grid(which='major', axis=opts['grid'], linestyle='--')
+
+    plt.yticks(opts['yticks'])
+
+    plt.title(opts['title'], fontsize = opts['titleSize'])
+    plt.tight_layout()
+    plt.savefig(outputFile)
+    if opts['show'] is True:
+        plt.show()
+
+if __name__ == '__main__':
+    mainexecution()