用于查询 MongoDB 集合的 ASP.NET Web API 端点

Question

我的数据库中有两个集合：

// UnifiedPost
public class UnifiedPost
{
    public UnifiedPost() { Id = ObjectId.GenerateNewId(); }

    [BsonId]
    public ObjectId Id { get; set; }

    public int OriginalId { get; set; }
    public string Thumbnail { get; set; }
    public string Preview { get; set; }
    [AlsoNotifyFor("IsVideo")]
    public string FullSized { get; set; }
    public char Rating { get; set; }
    public int Width { get; set; }
    public int Height { get; set; }
    public string Dimensions { get { return string.Format("{0}x{1}", Width, Height); } }
    public int Score { get; set; }
    [MongoDB.Bson.Serialization.Attributes.BsonIgnore]
    public string Tags { get; set; }
    public ICollection<ObjectId> TagIds { get; set; }
    public char Site { get; set; }

    public bool IsVideo { get { return FullSized.Contains(".mp4") || FullSized.Contains(".webm"); } }

    public string UniversalId { get; set; } // UNIQUE index (Site + Original id; fe. S3001)
}

// Tag
public class Tag
{
    public Tag() { }

    public ObjectId Id { get; set; }
    public string Name { get; set; }
}

我正在尝试创建一个 WebAPI 端点以从数据库中查询帖子：

[HttpGet, Route("post")]
public UnifiedPost[] GetPosts(string tags = null, string sites = "RSGLY", string rating = null,int page = 0, int limit = 100, string sort = "id:desc")
{
    if (limit == 0 || limit > 100) limit = 100;
    var toSkip = page * limit;
    var tagArray = tags.Split(new char[] { ' ' }, StringSplitOptions.RemoveEmptyEntries);
    var siteArray = sites.ToCharArray();
    var client = new MongoClient();
    var db = client.GetDatabase("Booru");
    var postsColl = db.GetCollection<UnifiedPost>("Posts");
    var tagsColl = db.GetCollection<BooruScraper.Models.Tag>("Tags");
    var tagIds = tagsColl.Find(o => tagArray.Contains(o.Name)).ToList().ToDictionary(o => o.Name, o => o.Id);

    var query = postsColl.Find<UnifiedPost>(o => siteArray.Contains(o.Site) && tagIds.Values.All(p => o.TagIds.Contains(p))).Project<UnifiedPost>(Builders<UnifiedPost>.Projection.Expression(p => new UnifiedPost()
    {
        FullSized = p.FullSized,
        Height = p.Height,
        Id = p.Id,
        OriginalId = p.OriginalId,
        Preview = p.Preview,
        Rating = p.Rating,
        Score = p.Score,
        Site = p.Site,
        TagIds = new ObjectId[0],
        Tags = string.Join(" ", tagsColl.Find(q => p.TagIds.Contains(q.Id), new FindOptions()).Project<string>(Builders<BooruScraper.Models.Tag>.Projection.Expression(r => r.Name)).ToList(CancellationToken.None)),
        Thumbnail = p.Thumbnail,
        UniversalId = p.UniversalId,
        Width = p.Width
    })).Skip(toSkip).ToList();
    return query.ToArray();
}

坦率地说，这段代码很丑陋，甚至没有达到最低目的。这简直太可悲了，但我想不出任何方法可以用 Find Fluent API 做到这一点。

一方面，我想按用户输入的标签过滤所有帖子，如果没有输入标签，则不应用过滤器。但我不知道如何实现这一点，更不用说如何将所有其他过滤器（站点、排序、页面...）应用于find 查询。

我在这里缺少什么？如果Find 在这种情况下我应该使用什么？

谢谢！

编辑：用最简单的术语来说，我该怎么做：

1. 从整个帖子集合中筛选出与输入标签匹配的帖子；
2. 从该结果进一步过滤到仅用户输入的站点；
3. 进一步过滤收藏并删除所有与输入评级不匹配的帖子；
4. 页面、限制等

Answer 1

天亮了。我们从哪里开始？ ;)

除非有特定原因需要两个单独的集合，否则我建议您摆脱“标签”集合。即使您需要它（例如，为了集中管理可用标签或其他东西），您仍然希望在 UnifiedPost 文档中复制标签信息。文档数据库并不是为了真正支持连接而构建的。

这是一个工作示例，应该让您了解如何构建它（我将您的实体剥离到相关部分）：

using MongoDB.Bson;
using MongoDB.Driver;
using System;
using System.Linq;

namespace ConsoleApp1
{
    public class UnifiedPost
    {
        public ObjectId Id { get; set; }
        public string[] Tags { get; set; }
        public char Site { get; set; }
        public int Score { get; set; }
    }

    public class Program
    {
        static void Main(string[] args)
        {
            var collection = new MongoClient().GetDatabase("test").GetCollection<UnifiedPost>("test");

            collection.InsertOne(new UnifiedPost() { Site = 'X', Tags = new[] { "a" } });
            collection.InsertOne(new UnifiedPost() { Site = 'Y', Tags = new[] { "a", "b" } });
            collection.InsertOne(new UnifiedPost() { Site = 'Z', Tags = new[] { "a", "b", "c" } });

            // 1. first filter
            var tagsWeAreInterestedIn = new[] { "a", "b" };
            var tagsFilter = Builders<UnifiedPost>.Filter.All(up => up.Tags, tagsWeAreInterestedIn);

            // 2. second filter
            var sitesWeAreInterestedIn = new[] { 'X', 'Y' };
            var siteFilter = Builders<UnifiedPost>.Filter.In(up => up.Site, sitesWeAreInterestedIn);

            // 3. third filter: rating (would follow the exact same pattern as the sites filter so we skip that part)

            // search for combined filter
            var query = collection.Find(tagsFilter & siteFilter);

            // 4. paging (e.g. sorted by the Score property, 10 results per page, skipped to page 4)
            // please note: this example will not return any data with skip=3
            // because we do not insert enough test data - use skip(0) to get a result!
            query = query.SortBy(up => up.Score).Skip(3).Limit(10);

            var postsWithTagB = query.ToList();

            Console.ReadLine();
        }
    }