opendal_service_webhdfs/
backend.rs

1// Licensed to the Apache Software Foundation (ASF) under one
2// or more contributor license agreements.  See the NOTICE file
3// distributed with this work for additional information
4// regarding copyright ownership.  The ASF licenses this file
5// to you under the Apache License, Version 2.0 (the
6// "License"); you may not use this file except in compliance
7// with the License.  You may obtain a copy of the License at
8//
9//   http://www.apache.org/licenses/LICENSE-2.0
10//
11// Unless required by applicable law or agreed to in writing,
12// software distributed under the License is distributed on an
13// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
14// KIND, either express or implied.  See the License for the
15// specific language governing permissions and limitations
16// under the License.
17
18use std::fmt::Debug;
19use std::sync::Arc;
20
21use asyncband::once::OnceCell;
22use bytes::Buf;
23use http::StatusCode;
24use log::debug;
25
26use super::WEBHDFS_SCHEME;
27use super::config::WebhdfsConfig;
28use super::core::parse_error;
29use super::core::{ErrorContext, WebhdfsCore};
30use super::deleter::WebhdfsDeleter;
31use super::lister::WebhdfsLister;
32use super::message::BooleanResp;
33use super::message::FileStatusType;
34use super::message::FileStatusWrapper;
35use super::reader::*;
36use super::writer::WebhdfsWriter;
37use super::writer::WebhdfsWriters;
38use opendal_core::raw::oio;
39use opendal_core::raw::*;
40use opendal_core::*;
41
42const WEBHDFS_DEFAULT_ENDPOINT: &str = "http://127.0.0.1:9870";
43
44/// [WebHDFS](https://hadoop.apache.org/docs/stable/hadoop-project-dist/hadoop-hdfs/WebHDFS.html)'s REST API support.
45#[doc = include_str!("docs.md")]
46#[derive(Debug, Default)]
47pub struct WebhdfsBuilder {
48    pub(super) config: WebhdfsConfig,
49}
50
51impl WebhdfsBuilder {
52    /// Set the working directory of this backend
53    ///
54    /// All operations will happen under this root
55    ///
56    /// # Note
57    ///
58    /// The root will be automatically created if not exists.
59    pub fn root(mut self, root: &str) -> Self {
60        self.config.root = if root.is_empty() {
61            None
62        } else {
63            Some(root.to_string())
64        };
65
66        self
67    }
68
69    /// Set the remote address of this backend
70    /// default to `http://127.0.0.1:9870`
71    ///
72    /// Endpoints should be full uri, e.g.
73    ///
74    /// - `https://webhdfs.example.com:9870`
75    /// - `http://192.168.66.88:9870`
76    ///
77    /// If user inputs endpoint without scheme, we will
78    /// prepend `http://` to it.
79    pub fn endpoint(mut self, endpoint: &str) -> Self {
80        if !endpoint.is_empty() {
81            // trim tailing slash so we can accept `http://127.0.0.1:9870/`
82            self.config.endpoint = Some(endpoint.trim_end_matches('/').to_string());
83        }
84        self
85    }
86
87    /// Set the username of this backend,
88    /// used for authentication
89    pub fn user_name(mut self, user_name: &str) -> Self {
90        if !user_name.is_empty() {
91            self.config.user_name = Some(user_name.to_string());
92        }
93        self
94    }
95
96    /// Set the delegation token of this backend,
97    /// used for authentication
98    ///
99    /// # Note
100    /// The builder prefers using delegation token over username.
101    /// If both are set, delegation token will be used.
102    pub fn delegation(mut self, delegation: &str) -> Self {
103        if !delegation.is_empty() {
104            self.config.delegation = Some(delegation.to_string());
105        }
106        self
107    }
108
109    /// Disable batch listing
110    ///
111    /// # Note
112    ///
113    /// When listing a directory, the backend will default to use batch listing.
114    /// If disabled, the backend will list all files/directories in one request.
115    pub fn disable_list_batch(mut self) -> Self {
116        self.config.disable_list_batch = true;
117        self
118    }
119
120    /// Set temp dir for atomic write.
121    ///
122    /// # Notes
123    ///
124    /// If not set, write multi not support, eg: `.opendal_tmp/`.
125    pub fn atomic_write_dir(mut self, dir: &str) -> Self {
126        self.config.atomic_write_dir = if dir.is_empty() {
127            None
128        } else {
129            Some(String::from(dir))
130        };
131        self
132    }
133}
134
135impl Builder for WebhdfsBuilder {
136    type Config = WebhdfsConfig;
137
138    /// build the backend
139    ///
140    /// # Note
141    ///
142    /// when building backend, the built backend will check if the root directory
143    /// exits.
144    /// if the directory does not exit, the directory will be automatically created
145    fn build(self) -> Result<impl Service> {
146        debug!("start building backend: {self:?}");
147
148        let root = normalize_root(&self.config.root.unwrap_or_default());
149        debug!("backend use root {root}");
150
151        // check scheme
152        let endpoint = match self.config.endpoint {
153            Some(endpoint) => {
154                if endpoint.starts_with("http") {
155                    endpoint
156                } else {
157                    format!("http://{endpoint}")
158                }
159            }
160            None => WEBHDFS_DEFAULT_ENDPOINT.to_string(),
161        };
162        debug!("backend use endpoint {endpoint}");
163
164        let atomic_write_dir = self.config.atomic_write_dir;
165
166        let auth = self.config.delegation.map(|dt| format!("delegation={dt}"));
167
168        let info = ServiceInfo::new(WEBHDFS_SCHEME, &root, "");
169        let capability = Capability {
170            stat: true,
171
172            read: true,
173
174            write: true,
175            write_can_append: true,
176            write_can_multi: atomic_write_dir.is_some(),
177
178            create_dir: true,
179            delete: true,
180
181            list: true,
182
183            shared: true,
184
185            ..Default::default()
186        };
187
188        let accessor_info = info;
189        let core = Arc::new(WebhdfsCore {
190            info: accessor_info,
191            capability,
192            root,
193            endpoint,
194            user_name: self.config.user_name,
195            auth,
196            root_checker: OnceCell::new(),
197            atomic_write_dir,
198            disable_list_batch: self.config.disable_list_batch,
199        });
200
201        Ok(WebhdfsBackend { core })
202    }
203}
204
205/// Backend for WebHDFS service
206#[derive(Debug, Clone)]
207pub struct WebhdfsBackend {
208    pub(crate) core: Arc<WebhdfsCore>,
209}
210
211impl WebhdfsBackend {
212    async fn check_root(&self, ctx: &OperationContext) -> Result<()> {
213        let resp = self.core.webhdfs_get_file_status(ctx, "/").await?;
214        match resp.status() {
215            StatusCode::OK => {
216                let bs = resp.into_body();
217
218                let file_status = serde_json::from_reader::<_, FileStatusWrapper>(bs.reader())
219                    .map_err(new_json_deserialize_error)?
220                    .file_status;
221
222                if file_status.ty == FileStatusType::File {
223                    return Err(Error::new(
224                        ErrorKind::ConfigInvalid,
225                        "root path must be dir",
226                    ));
227                }
228            }
229            StatusCode::NOT_FOUND => {
230                self.create_dir(ctx, "/", OpCreateDir::new()).await?;
231            }
232            _ => {
233                return Err(parse_error(
234                    ErrorContext::new(ServiceOperation("GetFileStatus")),
235                    resp,
236                ));
237            }
238        }
239        Ok(())
240    }
241}
242
243impl Service for WebhdfsBackend {
244    type Reader = oio::StreamReader<WebhdfsReader>;
245    type Writer = WebhdfsWriters;
246    type Lister = oio::PageLister<WebhdfsLister>;
247    type Deleter = oio::OneShotDeleter<WebhdfsDeleter>;
248    type Copier = ();
249    type Composer = ();
250
251    fn info(&self) -> ServiceInfo {
252        self.core.info.clone()
253    }
254
255    fn capability(&self) -> Capability {
256        self.core.capability
257    }
258
259    /// Create a file or directory
260    async fn create_dir(
261        &self,
262        ctx: &OperationContext,
263        path: &str,
264        _: OpCreateDir,
265    ) -> Result<RpCreateDir> {
266        let resp = self.core.webhdfs_create_dir(ctx, path).await?;
267
268        let status = resp.status();
269        // WebHDFS's has a two-step create/append to prevent clients to send out
270        // data before creating it.
271        // According to the redirect policy of `reqwest` HTTP Client we are using,
272        // the redirection should be done automatically.
273        match status {
274            StatusCode::CREATED | StatusCode::OK => {
275                let bs = resp.into_body();
276
277                let resp = serde_json::from_reader::<_, BooleanResp>(bs.reader())
278                    .map_err(new_json_deserialize_error)?;
279
280                if resp.boolean {
281                    Ok(RpCreateDir::default())
282                } else {
283                    Err(Error::new(
284                        ErrorKind::Unexpected,
285                        "webhdfs create dir failed",
286                    ))
287                }
288            }
289            _ => Err(parse_error(
290                ErrorContext::new(ServiceOperation("Mkdirs")),
291                resp,
292            )),
293        }
294    }
295
296    async fn stat(&self, ctx: &OperationContext, path: &str, _: OpStat) -> Result<RpStat> {
297        // if root exists and is a directory, stat will be ok
298        self.core
299            .root_checker
300            .get_or_try_init(|| async { self.check_root(ctx).await })
301            .await?;
302
303        let resp = self.core.webhdfs_get_file_status(ctx, path).await?;
304        let status = resp.status();
305        match status {
306            StatusCode::OK => {
307                let bs = resp.into_body();
308
309                let file_status = serde_json::from_reader::<_, FileStatusWrapper>(bs.reader())
310                    .map_err(new_json_deserialize_error)?
311                    .file_status;
312
313                let meta = match file_status.ty {
314                    FileStatusType::Directory => MetadataBuilder::dir().build(),
315                    FileStatusType::File => {
316                        let mut metadata = MetadataBuilder::file(file_status.length);
317                        metadata.last_modified(Timestamp::from_millisecond(
318                            file_status.modification_time,
319                        )?);
320                        metadata.build()
321                    }
322                };
323
324                Ok(RpStat::new(meta))
325            }
326
327            _ => Err(parse_error(
328                ErrorContext::new(ServiceOperation("GetFileStatus")),
329                resp,
330            )),
331        }
332    }
333    fn read(&self, ctx: &OperationContext, path: &str, args: OpRead) -> Result<Self::Reader> {
334        let output: oio::StreamReader<WebhdfsReader> = {
335            Ok(oio::StreamReader::new(WebhdfsReader::new(
336                self.clone(),
337                ctx.clone(),
338                path,
339                args,
340            )))
341        }?;
342
343        Ok(output)
344    }
345
346    fn write(&self, ctx: &OperationContext, path: &str, args: OpWrite) -> Result<Self::Writer> {
347        let output: WebhdfsWriters = {
348            let w = WebhdfsWriter::new(
349                self.core.clone(),
350                ctx.clone(),
351                args.clone(),
352                path.to_string(),
353            );
354
355            let w = if args.append() {
356                WebhdfsWriters::Two(oio::AppendWriter::new(w))
357            } else {
358                WebhdfsWriters::One(oio::BlockWriter::new(
359                    ctx.executor().clone(),
360                    w,
361                    args.concurrent(),
362                ))
363            };
364
365            Ok(w)
366        }?;
367
368        Ok(output)
369    }
370
371    fn delete(&self, ctx: &OperationContext) -> Result<Self::Deleter> {
372        let output: oio::OneShotDeleter<WebhdfsDeleter> = {
373            Ok(oio::OneShotDeleter::new(WebhdfsDeleter::new(
374                self.core.clone(),
375                ctx.clone(),
376            )))
377        }?;
378
379        Ok(output)
380    }
381
382    fn list(&self, ctx: &OperationContext, path: &str, args: OpList) -> Result<Self::Lister> {
383        let output: oio::PageLister<WebhdfsLister> = {
384            if args.recursive() {
385                return Err(Error::new(
386                    ErrorKind::Unsupported,
387                    "WebHDFS doesn't support list with recursive",
388                ));
389            }
390
391            let path = path.trim_end_matches('/');
392            let l = WebhdfsLister::new(self.core.clone(), ctx.clone(), path);
393            Ok(oio::PageLister::new(l))
394        }?;
395
396        Ok(output)
397    }
398
399    fn copy(
400        &self,
401        _ctx: &OperationContext,
402        _from: &str,
403        _to: &str,
404        _args: OpCopy,
405    ) -> Result<Self::Copier> {
406        Err(Error::new(
407            ErrorKind::Unsupported,
408            "operation is not supported",
409        ))
410    }
411
412    async fn rename(
413        &self,
414        _ctx: &OperationContext,
415        _from: &str,
416        _to: &str,
417        _args: OpRename,
418    ) -> Result<RpRename> {
419        Err(Error::new(
420            ErrorKind::Unsupported,
421            "operation is not supported",
422        ))
423    }
424
425    async fn presign(
426        &self,
427        _ctx: &OperationContext,
428        _path: &str,
429        _args: OpPresign,
430    ) -> Result<RpPresign> {
431        Err(Error::new(
432            ErrorKind::Unsupported,
433            "operation is not supported",
434        ))
435    }
436}
opendal_service_webhdfs/backend.rs

opendal_service_webhdfs/
backend.rs
