Azure functions 在Azure函数中加载第三方dll

我在Azure函数中使用2.1运行时。当我构建icu dll时（

icui59.dll

，

icuio59.dll

，

icutest59.dll

，

icutu59.dll

，

icuuc59.dll

，以及

icudt59.dll

）被复制到

。/netcoreapp2.1/lib/win-x64

文件夹以及

。/netcoreapp2.1/bin/runtimes/win7-x64/native>文件夹
icu.net.dll
也被复制到
。/netcoreapp2.1/bin

我在Windows 10上本地运行

当我的函数执行时，我得到：

[20/03/2019 16:09:14] Executed 'AnalyzeVocabulary' (Failed, Id=cb768548-dd06-4f95-b889-af244af716d1)
[20/03/2019 16:09:14] System.Private.CoreLib: Exception while executing function: AnalyzeVocabulary. icu.net: Can't load ICU library (version 0).
[20/03/2019 16:09:14] 926e7209782c408984ad827cc2744b2d: Function 'AnalyzeVocabulary (Activity)' failed with an error. Reason: System.IO.FileLoadException: Can't load ICU library (version 0)
[20/03/2019 16:09:14] File name: 'icuuc'
[20/03/2019 16:09:14]    at Icu.NativeMethods.LoadIcuLibrary(String libraryName)
[20/03/2019 16:09:14]    at Icu.NativeMethods.get_IcuCommonLibHandle()
[20/03/2019 16:09:14]    at Icu.NativeMethods.uloc_canonicalize(String localeID, IntPtr name, Int32 nameCapacity, ErrorCode& err)
[20/03/2019 16:09:14]    at Icu.Locale.<>c__DisplayClass44_0`1.<GetString>b__0(IntPtr ptr, Int32 length)
[20/03/2019 16:09:14]    at Icu.NativeMethods.GetString(Func`3 lambda, Boolean isUnicodeString, Int32 initialLength)
[20/03/2019 16:09:14]    at Icu.Locale.GetString[T](GetStringMethod`1 method, T localeId)
[20/03/2019 16:09:14]    at Icu.Locale.Canonicalize(String localeID)
[20/03/2019 16:09:14]    at Icu.Locale.op_Implicit(String localeId)
[20/03/2019 16:09:14]    at DigitalLinguistics.TextAnalytics.TextSegApiTokenizer.TextTokenizer.SegmentSentences(String sentence) in E:\DigitalLinguistics\Dev\ServerlessSentinel\DigitalLinguistics.TextAnalytics\TextTokenizer.cs:line 70
[20/03/2019 16:09:14]    at DigitalLinguistics.TextAnalytics.TextSegApiTokenizer.TextTokenizer.Tokenize(String sentence, String id) in E:\DigitalLinguistics\Dev\ServerlessSentinel\DigitalLinguistics.TextAnalytics\TextTokenizer.cs:line 82
[20/03/2019 16:09:14]    at DigitalLinguistics.TextAnalytics.DataStructures.Sentence.LineParse(String text, String id, TokenizerOptions tokenOpts) in E:\DigitalLinguistics\Dev\ServerlessSentinel\DigitalLinguistics.TextAnalytics\Sentence.cs:line 92

[20/03/2019 16:09:14]执行“分析词汇表”（失败，Id=cb768548-dd06-4f95-b889-af244af716d1）
[20/03/2019 16:09:14]System.Private.CoreLib:执行函数时出现异常：Analyze词汇表。icu.net:无法加载icu库（版本0）。
[20/03/2019 16:09:14]926e7209782c408984ad827cc2744b2d:函数“分析词汇表（活动）”因错误而失败。原因：System.IO.FileLoadException:无法加载ICU库（版本0）
[20/03/2019 16:09:14]文件名：“icuuc”
[20/03/2019 16:09:14]在Icu.NativeMethods.LoadicLibrary（字符串库名）
[20/03/2019 16:09:14]在重症监护病房。本地方法。获取
[20/03/2019 16:09:14]在Icu.NativeMethods.uloc_规范化（字符串localeID、IntPtr名称、Int32名称容量、错误代码和错误）
[20/03/2019 16:09:14]在Icu.Locale.c_uuuu显示类44_u0`1.b_uuu0（IntPtr ptr，Int32长度）
[20/03/2019 16:09:14]在Icu.NativeMethods.GetString（Func`3 lambda，Boolean isUnicodeString，Int32 initialLength）
[20/03/2019 16:09:14]在Icu.Locale.GetString[T]（GetStringMethod`1方法，T localeId）
[20/03/2019 16:09:14]在Icu.Locale.Canonicalize（字符串localeID）
[20/03/2019 16:09:14]在Icu.Locale.op_隐式（字符串localeId）
[20/03/2019 16:09:14]在数字语言学.TextAnalytics.TextSegApiTokenizer.TextTokenizer.SegmentSequences（字符串句子）中，E:\DigitalLanguagements\Dev\serverlessentennel\DigitalLanguagements.TextAnalytics\TextTokenizer.cs:第70行
[20/03/2019 16:09:14]在E:\DigitalLanguagements\Dev\serverlessentennel\digitalLanguagements.TextAnalytics\TextTokenizer.TextTokenizer.Tokenize（字符串句子，字符串id）中的数字语言学.TextAnalytics\TextTokenizer.cs:第82行
[20/03/2019 16:09:14]在E:\DigitalLanguagements\Dev\serverlessentennel\DigitalLanguagements.TextAnalytics\句子中的DigitalLanguagements.DataStructures.句子.LineParse（字符串文本、字符串id、标记选项标记选项）中

注意我已经成功地从net core控制台应用程序运行了这些库，所以我猜这些库不在正确的文件夹中

关于这些库需要在哪里以及如何设置它们，以便将它们部署并发布在那里，作为我使用VS 2017构建的一部分，有什么建议吗？
我从GitHub中的源代码中发现，位置必须是
。/netcoreapp2.1/bin/lib