homework for it-support-engineer Bruce Diao

it-support-engineer/InterviewLog.ps1

@@ -0,0 +1,131 @@
+<#
+    将输入的原始日志时间格式转换为timeWindow格式，并返回一个timeWindow的字符串
+#>
+Function ConvertTo-TimeWindow
+{
+    [CmdletBinding()]
+    Param
+    (
+        [Parameter(Mandatory = $true, Position = 0)]$TimeString
+    )
+    [int]$Hour = ($TimeString -split":")[0]
+
+    switch ($Hour)
+    {
+        {$_ -lt 9} {[string]$StartTimeWindow = '0' + $Hour; [string]$EndTimeWindow = '0' + ($Hour + 1);[string]$TimeWindowString = $StartTimeWindow +"00" + "-" + $EndTimeWindow + "00";return $TimeWindowString}
+        {$_ -eq 9} {[string]$StartTimeWindow = '0' + $Hour; [string]$TimeWindowString = $StartTimeWindow +"00" + "-" + "1000";return $TimeWindowString}
+        {$_ -ge 10} {[string]$StartTimeWindow = $Hour;[string]$EndTimeWindow = $Hour + 1;[string]$TimeWindowString = $StartTimeWindow +"00"+ "-" + $EndTimeWindow + "00";return $TimeWindowString}
+    }
+
+}
+
+<#
+    将原始日志文件根据条件筛选后进行对应字符串处理，并格式化timeWindow日期，返回一个数组对象
+#>
+Function ConvertTo-LogFile
+{
+    [CmdletBinding()]
+    Param
+    (
+        [Parameter(Mandatory = $true, Position = 0)]$LogPath
+    )
+    $logs = Get-Content $LogPath
+    $Type = @()
+    $LogCount = $logs.Count
+    for ($i = 0; $i -lt $LogCount ; $i++)
+    { 
+        switch ($logs[$i])
+        {
+            {($_ -match '.*\[\d.*\]') -and ($_ -notmatch 'Notice\:') -and ($_ -notmatch 'WARNING\:') -and ($_ -notmatch 'failed\:')} {
+                $Type += New-Object -TypeName Psobject -Property @{
+                    "deviceName" = ($_ -split " ",6)[3]
+                    "processId" = ($_ -split " ",6)[4].Split("[")[1].Split("]")[0]
+                    "processName" = ($_ -split " ",6)[4].Split("[")[0]
+                    "description"=  ($_ -split " ",6)[5]
+                    "timeWindow" = ConvertTo-TimeWindow -TimeString ($_ -split " ",4)[2].Split(":")[0]
+                    }
+                }
+
+            {$_ -match 'Configuration Notice\:'} {
+                $Type += New-Object -TypeName Psobject -Property @{
+                    "deviceName" = ($_ -split " ",6)[3]
+                    "processId" = ($_ -split " ",6)[4].Split("[")[1].Split("]")[0]
+                    "processName" = ($_ -split " ",6)[4].Split("[")[0]
+                    "description"= ($_ -split ":",4)[3] + $logs[$i+1] + $logs[$i+2]
+                    "timeWindow" = ConvertTo-TimeWindow -TimeString ($_ -split " ",4)[2].Split(":")[0]
+                    }
+                }
+
+            {($_ -match 'WARNING\:') -or ($_ -match 'failed\:')} {
+                $Type += New-Object -TypeName Psobject -Property @{
+                    "deviceName" = ($_ -split " ",6)[3]
+                    "processId" = ($_ -split ":",8)[2].Split("[")[1].Split("]")[0]
+                    "processName" = ($_ -split " ",5)[4].Split(":")[0].Split("[")[0]
+                    "description"= ($_ -split ":",4)[3]
+                    "timeWindow" = ConvertTo-TimeWindow -TimeString ($_ -split " ",4)[2].Split(":")[0]
+                    }
+                }
+        }
+    }
+    return $Type
+}
+
+<#
+    输入原始日志文件及需要提交的网址，最终生成指定Key的Josn文件并上传到服务器�?
+#>
+
+Function Submit-PostRequest
+{
+    [CmdletBinding()]
+    Param
+    (
+        [Parameter(Mandatory = $true, Position = 0)]$FilePath,
+        [Parameter(Mandatory = $true, Position = 1)]$Url,
+        [switch]$LogCsv 
+    )
+# 将原始数据按条件进行字符串处理
+    $LogFile = ConvertTo-LogFile -LogPath $FilePath
+
+# 将日志按小时级内发生的次数进行处理（去除重复日志信息）
+    $JosnReport = @()
+    $LogFile | Group-Object -Property  timeWindow, deviceName, processId, processName, description | ForEach-Object{
+            $JosnReport += New-Object -TypeName PsObject -Property @{
+            "deviceName"  = $_.Group[0].deviceName
+            "processId"   = $_.Group[0].processId
+            "processName" = $_.Group[0].processName
+            "description" = $_.Group[0].description
+            "timeWindow"  = $_.Group[0].timeWindow
+            "numberOfOccurrence" = $_.Count
+        }
+    }
+# 将数组对象转换为Json格式
+    $SubmitJson = $JosnReport | Select-Object -Property deviceName,processId,processName,description,timeWindow,numberOfOccurrence | ConvertTo-Json
+
+# 如果有LogCsv开关参数打开，输出CSV格式日志到本地
+if ($LogCsv) {
+    $JosnReport | Export-Csv -Path .\Interview.csv -NoTypeInformation -Encoding UTF8
+}
+
+# Windows10、11 默认Powershell版本5.1，用于不受信任证书的运行环境
+
+add-type @"
+using System.Net;
+using System.Security.Cryptography.X509Certificates;
+public class TrustAllCertsPolicy : ICertificatePolicy {
+    public bool CheckValidationResult(
+        ServicePoint srvPoint, X509Certificate certificate,
+        WebRequest request, int certificateProblem) {
+            return true;
+        }
+ }
+"@
+[System.Net.ServicePointManager]::CertificatePolicy = New-Object TrustAllCertsPolicy
+
+    Invoke-WebRequest -ContentType "application/json" -Method POST -Body $SubmitJson -Uri $url
+}
+
+
+<#
+    执行主函数，输入原始日志文件及需要提交的网址,需要本地输出可以添加“-Logcsv”开关参数
+#>
+Submit-PostRequest -FilePath "D:\interview_data_set" -Url "https://foo.com/bar"

it-support-engineer/InterviewLog.py

@@ -0,0 +1,85 @@
+#!/usr/bin/env python3
+import re
+import json
+import requests
+
+def main(logfile,url):
+    # 将原始日志进行字符串处理
+    json_dict=list_logs(logfile)  
+    # 去除重复日志
+    new_list=group(json_dict)
+    # 统计日志发生次数
+    new_list=count(json_dict,new_list)
+    # Post Json文件到服务器
+    send_json(new_list,url) 
+    # print(json_dict)
+
+def list_logs(filepath):   
+    json_dict=[]
+    with open(filepath) as file:
+        lines= file.readlines()
+    for text in lines:     
+        if '---' in text:     
+            text=pre_text
+        elif 'Notice:' in text:    
+            t=0
+            pre_text=text
+            continue
+        elif '	' in text and t!=1 :  
+            pre_text=pre_text.split('\n')[0]+text.split('	')[1]
+            t+=1
+            continue
+        elif '	' in text and t==1:
+            text=pre_text+text        
+        json_dict.append(analysis(text))
+        pre_text=text
+    return json_dict
+
+def group(list):   
+    new_list = []
+    for it in list:
+        if it not in new_list:
+            new_list.append(it)
+    return new_list
+
+def count(list,new_list): 
+    # print(new_list)
+        for it in new_list:
+            total=list.count(it)
+            it['numberOfOccurrence']=total
+            # if total>1:
+            #     print(it)
+        return new_list
+
+def send_json(list,url): 
+        requests.packages.urllib3.disable_warnings()
+        headers={'Content-Type': 'application/json;charset=utf-8'}
+        try:
+            response = requests.post(url=url, headers=headers, data=json.dumps(list),verify=False)
+            if response.status_code == 200:
+                # return response.json()
+                print(response)
+            else:
+                print(response)
+        except requests.ConnectionError as e:
+            print('Error', e.args)
+
+def analysis(dump):
+    data={}
+    data['deviceName']=dump.split()[3]
+    data['processName']=re.match('^[a-zA-Z\.]*',dump.split()[4]).group()
+    data['processId'] = dump.split('[')[1].split(']')[0]
+    data['description']=' '.join(dump.split()[5:-1])
+    hour=re.match('^([0-9]{2})',dump.split()[2]).group()
+    if int(hour)<9:
+        data['timeWindow']='0'+hour[1]+'00-0'+str(int(hour)+1)+'00'
+    elif int(hour)==9:
+        data['timeWindow'] ='0'+hour + '00-' + str(int(hour) + 1) + '00'
+    else:
+        data['timeWindow'] = hour + '00-' + str(int(hour) + 1) + '00'
+    return data
+
+if __name__=='__main__':
+    logpath='.\interview_data_set'
+    url='https://foo.com/bar'
+    main(logpath,url)