public IEnumerable <T> ParseInput <T>(string input)
{
var xml = XElement.Parse(input);
// some more code here
var parser = _parserFactory.CreateParser <T>();
return(parser.Parse(xml));
}
protected virtual void VisitSolution(string solutionPath)
{
var parser = _solutionParserFactory.CreateParser(solutionPath);
var solution = parser.Parse(solutionPath);
_codeBase.Solutions.Add(solution);
}
protected virtual void VisitProject(string projectPath)
{
var parser = _projectParserFactory.CreateParser(projectPath);
var project = parser.Parse(projectPath);
project.Repository = _repositoryRegistry.Current;
_codeBase.Projects.Add(project);
}
public async Task <IEnumerable <IWorkTask> > WorkAsync(IDownloader downloader, IParserFactory parserFactory)
{
var html = await downloader.Down(url);
var parser = parserFactory.CreateParser("xpath");
parser.Init(html);
var keyValuePairs = parser.Parser(new ParaseRule[] {
new ParaseRule()
{
Name = "博客列表",
Path = "//div[@class=\"post_item\"]",
Rules = new ParaseRule[] {
new ParaseRule()
{
Path = "*//a[@class=\"titlelnk\"]",
Name = "标题"
},
new ParaseRule()
{
Path = "*//a[@class=\"titlelnk\"]",
Name = "链接",
AttrName = "href"
}
}
}
});
foreach (var item in keyValuePairs)
{
Console.WriteLine($"{item}");
}
var nextUrlKv = parser.Parser(new ParaseRule[] {
new ParaseRule()
{
Name = "下一页",
Path = "//div[@class=\"pager\"]/a[contains(@class,'current')]/following-sibling::a[1]",
AttrName = "href"
}
});
var results = nextUrlKv.Select(x => (IWorkTask) new SpiderTask(x.Value));
return(await Task.FromResult(results));
}
/// <summary>
/// Standard implementation
/// </summary>
/// <returns>Consolidated schedules</returns>
public IEnumerable <IRecord> Read()
{
var parser = _cifParserFactory.CreateParser();
return(parser.Read(Extract()));
}
public IParser CreateParser()
{
return(new ScheduleConsolidator(_factory.CreateParser(), _logger));
}