c# - yield 模式，状态机流程

Question

我有以下文件，我正在使用迭代器 block 来解析文件中某些重复出现的节点/部分。我最初使用正则表达式来解析整个文件，但是当节点中不存在某些字段时，它不会匹配。所以我正在尝试使用 yield 模式。文件格式如下，我正在使用的代码。我想要从文件中得到的只是将复制节点作为一个单独的部分，这样我就可以使用键字符串获取其中的字段并存储在对象集合中。我可以从第一个复制发生的地方开始解析，但无法在复制节点结束的地方结束它。

文件格式:

X_HEADER
{
    DATA_MANAGEMENT_FIELD_2     NA
    DATA_MANAGEMENT_FIELD_3     NA
    DATA_MANAGEMENT_FIELD_4     NA
    SYSTEM_SOFTWARE_VERSION     NA
}
Y_HEADER
{
    DATA_MANAGEMENT_FIELD_2     NA
    DATA_MANAGEMENT_FIELD_3     NA
    DATA_MANAGEMENT_FIELD_4     NA
    SYSTEM_SOFTWARE_VERSION     NA
}
COMPLETION
{
    NUMBER          877
    VERSION         4
    CALIBRATION_VERSION 1
    CONFIGURATION_ID    877    
}
REPLICATE
{
    REPLICATE_ID            1985
    ASSAY_NUMBER            656
    ASSAY_VERSION           4
    ASSAY_STATUS            Research
    DILUTION_ID         1
}
REPLICATE
{
    REPLICATE_ID            1985
    ASSAY_NUMBER            656
    ASSAY_VERSION           4
    ASSAY_STATUS            Research
}

代码:

static IEnumerable<IDictionary<string, string>> ReadParts(string path)
{
    using (var reader = File.OpenText(path))
    {
        var current = new Dictionary<string, string>();
        string line;
        while ((line = reader.ReadLine()) != null)
        {
            if (string.IsNullOrWhiteSpace(line)) continue;

            if (line.StartsWith("REPLICATE"))
            {
                yield return current;
                current = new Dictionary<string, string>();
            }
            else
            {
                var parts = line.Split('\t');
            }

            if (current.Count > 0) yield return current;
        }
    }
}

public static void parseFile(string fileName)
    {
        foreach (var part in ReadParts(fileName))
        {
           //part["fIELD1"] will retireve certain values from the REPLICATE PART HERE
        }
    }

Answer 1

嗯，这听起来好像您只需要在获得右大括号时“关闭”一个部分，并且此时仅yield return。例如:

static IEnumerable<IDictionary<string, string>> ReadParts(string path)
{
    using (var reader = File.OpenText(path))
    {
        string currentName = null;
        IDictionary<string, string> currentMap = null;
        while ((line = reader.ReadLine()) != null)
        {
            if (string.IsNullOrWhiteSpace(line))
            {
                continue;
            }
            if (line == "{")
            {
                if (currentName == null || currentMap != null)
                {
                    throw new BadDataException("Open brace at wrong place");
                }
                currentMap = new Dictionary<string, string>();
            }
            else if (line == "}")
            {
                if (currentName == null || currentMap == null)
                {
                    throw new BadDataException("Closing brace at wrong place");
                }
                // Isolate the "REPLICATE-only" requirement to a single
                // line - if you ever need other bits, you can change this.
                if (currentName == "REPLICATE")
                {
                    yield return currentMap;
                }
                currentName = null;
                currentMap = null;
            }
            else if (!line.StartsWith("\t"))
            {
                if (currentName != null || currentMap != null)
                {
                    throw new BadDataException("Section name at wrong place");
                }
                currentName = line;
            }
            else
            {
                if (currentName == null || currentMap == null)
                {
                    throw new BadDataException("Name/value pair at wrong place");
                }
                var parts = line.Substring(1).Split('\t');
                if (parts.Length != 2)
                {
                    throw new BadDataException("Invalid name/value pair");
                }
                currentMap[parts[0]] = parts[1];
            }                
        }
    }
}

老实说，这是一个相当可怕的功能。我怀疑我会把它放在它自己的类中(可能是一个嵌套的类)来存储状态，并使每个处理程序都有自己的方法。哎呀，这实际上是状态模式有意义的情况:)