risingwave_connector/
with_options.rs

1// Copyright 2023 RisingWave Labs
2//
3// Licensed under the Apache License, Version 2.0 (the "License");
4// you may not use this file except in compliance with the License.
5// You may obtain a copy of the License at
6//
7//     http://www.apache.org/licenses/LICENSE-2.0
8//
9// Unless required by applicable law or agreed to in writing, software
10// distributed under the License is distributed on an "AS IS" BASIS,
11// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12// See the License for the specific language governing permissions and
13// limitations under the License.
14
15use std::collections::{BTreeMap, BTreeSet, HashMap};
16use std::marker::PhantomData;
17use std::time::Duration;
18
19use risingwave_pb::id::SecretId;
20use risingwave_pb::secret::PbSecretRef;
21
22use crate::error::ConnectorResult;
23use crate::sink::catalog::SinkFormatDesc;
24use crate::source::cdc::MYSQL_CDC_CONNECTOR;
25use crate::source::cdc::external::ExternalCdcTableType;
26use crate::source::iceberg::ICEBERG_CONNECTOR;
27use crate::source::{
28    ADBC_SNOWFLAKE_CONNECTOR, AZBLOB_CONNECTOR, BATCH_POSIX_FS_CONNECTOR, GCS_CONNECTOR,
29    KAFKA_CONNECTOR, LEGACY_S3_CONNECTOR, OPENDAL_S3_CONNECTOR, POSIX_FS_CONNECTOR,
30    PULSAR_CONNECTOR, UPSTREAM_SOURCE_KEY,
31};
32
33/// Marker trait for `WITH` options. Only for `#[derive(WithOptions)]`, should not be used manually.
34///
35/// This is used to ensure the `WITH` options types have reasonable structure.
36///
37/// TODO: add this bound for sink. There's a `SourceProperties` trait for sources, but no similar
38/// things for sinks.
39pub trait WithOptions {
40    #[doc(hidden)]
41    #[inline(always)]
42    fn assert_receiver_is_with_options(&self) {}
43}
44
45// Currently CDC properties are handled specially.
46// - It simply passes HashMap to Java DBZ.
47// - It's not handled by serde.
48// - It contains fields other than WITH options.
49// TODO: remove the workaround here. And also use #[derive] for it.
50
51impl<T: crate::source::cdc::CdcSourceTypeTrait> WithOptions
52    for crate::source::cdc::CdcProperties<T>
53{
54}
55
56// impl the trait for value types
57
58impl<T: WithOptions> WithOptions for Option<T> {}
59impl WithOptions for Vec<String> {}
60impl WithOptions for Vec<u64> {}
61impl WithOptions for HashMap<String, String> {}
62impl WithOptions for BTreeMap<String, String> {}
63
64impl WithOptions for String {}
65impl WithOptions for bool {}
66impl WithOptions for usize {}
67impl WithOptions for u8 {}
68impl WithOptions for u16 {}
69impl WithOptions for u32 {}
70impl WithOptions for u64 {}
71impl WithOptions for i32 {}
72impl WithOptions for i64 {}
73impl WithOptions for f64 {}
74impl WithOptions for std::time::Duration {}
75impl WithOptions for crate::connector_common::MqttQualityOfService {}
76impl WithOptions for crate::connector_common::SslMode {}
77impl WithOptions for crate::sink::file_sink::opendal_sink::PathPartitionPrefix {}
78impl WithOptions for crate::sink::kafka::CompressionCodec {}
79impl WithOptions for crate::sink::pulsar::PulsarRoutingMode {}
80impl WithOptions for crate::source::filesystem::file_common::CompressionFormat {}
81impl WithOptions for nexmark::config::RateShape {}
82impl WithOptions for nexmark::event::EventType {}
83impl<T> WithOptions for PhantomData<T> {}
84
85pub trait Get {
86    fn get(&self, key: &str) -> Option<&String>;
87}
88
89pub trait GetKeyIter {
90    fn key_iter(&self) -> impl Iterator<Item = &str>;
91}
92
93impl GetKeyIter for HashMap<String, String> {
94    fn key_iter(&self) -> impl Iterator<Item = &str> {
95        self.keys().map(|s| s.as_str())
96    }
97}
98
99impl Get for HashMap<String, String> {
100    fn get(&self, key: &str) -> Option<&String> {
101        self.get(key)
102    }
103}
104
105impl Get for BTreeMap<String, String> {
106    fn get(&self, key: &str) -> Option<&String> {
107        self.get(key)
108    }
109}
110
111impl GetKeyIter for BTreeMap<String, String> {
112    fn key_iter(&self) -> impl Iterator<Item = &str> {
113        self.keys().map(|s| s.as_str())
114    }
115}
116
117/// Utility methods for `WITH` properties (`HashMap` and `BTreeMap`).
118pub trait WithPropertiesExt: Get + GetKeyIter + Sized {
119    #[inline(always)]
120    fn get_connector(&self) -> Option<String> {
121        self.get(UPSTREAM_SOURCE_KEY).map(|s| s.to_lowercase())
122    }
123
124    #[inline(always)]
125    fn is_kafka_connector(&self) -> bool {
126        let Some(connector) = self.get_connector() else {
127            return false;
128        };
129        connector == KAFKA_CONNECTOR
130    }
131
132    #[inline(always)]
133    fn is_pulsar_connector(&self) -> bool {
134        let Some(connector) = self.get_connector() else {
135            return false;
136        };
137        connector == PULSAR_CONNECTOR
138    }
139
140    #[inline(always)]
141    fn is_mysql_cdc_connector(&self) -> bool {
142        let Some(connector) = self.get_connector() else {
143            return false;
144        };
145        connector == MYSQL_CDC_CONNECTOR
146    }
147
148    #[inline(always)]
149    fn get_sync_call_timeout(&self) -> Option<Duration> {
150        const SYNC_CALL_TIMEOUT_KEY: &str = "properties.sync.call.timeout"; // only from kafka props, add more if needed
151        self.get(SYNC_CALL_TIMEOUT_KEY)
152            // ignore the error is ok here, because we will parse the field again when building the properties and has more precise error message
153            .and_then(|s| duration_str::parse_std(s).ok())
154    }
155
156    #[inline(always)]
157    fn is_cdc_connector(&self) -> bool {
158        let Some(connector) = self.get_connector() else {
159            return false;
160        };
161        connector.contains("-cdc")
162    }
163
164    /// It is shared when `CREATE SOURCE`, and not shared when `CREATE TABLE`. So called "shareable".
165    fn is_shareable_cdc_connector(&self) -> bool {
166        self.is_cdc_connector() && ExternalCdcTableType::from_properties(self).can_backfill()
167    }
168
169    fn enable_transaction_metadata(&self) -> bool {
170        ExternalCdcTableType::from_properties(self).enable_transaction_metadata()
171    }
172
173    fn is_shareable_non_cdc_connector(&self) -> bool {
174        self.is_kafka_connector()
175    }
176
177    #[inline(always)]
178    fn is_iceberg_connector(&self) -> bool {
179        let Some(connector) = self.get_connector() else {
180            return false;
181        };
182        connector == ICEBERG_CONNECTOR
183    }
184
185    fn connector_need_pk(&self) -> bool {
186        // Currently only iceberg connector doesn't need primary key
187        // introduced in https://github.com/risingwavelabs/risingwave/pull/14971
188        // XXX: This seems not the correct way. Iceberg doesn't necessarily lack a PK.
189        // "batch source" doesn't need a PK?
190        // For streaming, if it has a PK, do we want to use it? It seems not safe.
191        !self.is_iceberg_connector()
192    }
193
194    fn is_legacy_fs_connector(&self) -> bool {
195        self.get(UPSTREAM_SOURCE_KEY)
196            .map(|s| s.eq_ignore_ascii_case(LEGACY_S3_CONNECTOR))
197            .unwrap_or(false)
198    }
199
200    fn is_new_fs_connector(&self) -> bool {
201        self.get(UPSTREAM_SOURCE_KEY)
202            .map(|s| {
203                s.eq_ignore_ascii_case(OPENDAL_S3_CONNECTOR)
204                    || s.eq_ignore_ascii_case(POSIX_FS_CONNECTOR)
205                    || s.eq_ignore_ascii_case(GCS_CONNECTOR)
206                    || s.eq_ignore_ascii_case(AZBLOB_CONNECTOR)
207            })
208            .unwrap_or(false)
209    }
210
211    fn is_batch_connector(&self) -> bool {
212        self.get(UPSTREAM_SOURCE_KEY)
213            .map(|s| {
214                s.eq_ignore_ascii_case(BATCH_POSIX_FS_CONNECTOR)
215                    || s.eq_ignore_ascii_case(ADBC_SNOWFLAKE_CONNECTOR)
216            })
217            .unwrap_or(false)
218    }
219
220    fn supports_full_reload_refresh(&self) -> bool {
221        self.get(UPSTREAM_SOURCE_KEY)
222            .map(|s| {
223                s.eq_ignore_ascii_case(OPENDAL_S3_CONNECTOR)
224                    || s.eq_ignore_ascii_case(GCS_CONNECTOR)
225                    || s.eq_ignore_ascii_case(BATCH_POSIX_FS_CONNECTOR)
226                    || s.eq_ignore_ascii_case(ICEBERG_CONNECTOR)
227                    || s.eq_ignore_ascii_case(ADBC_SNOWFLAKE_CONNECTOR)
228            })
229            .unwrap_or(false)
230    }
231
232    fn requires_singleton(&self) -> bool {
233        self.is_new_fs_connector() || self.is_iceberg_connector() || self.is_batch_connector()
234    }
235}
236
237impl<T: Get + GetKeyIter> WithPropertiesExt for T {}
238
239/// Options or properties extracted from the `WITH` clause of DDLs.
240#[derive(Default, Clone, Debug, PartialEq, Eq, Hash)]
241pub struct WithOptionsSecResolved {
242    inner: BTreeMap<String, String>,
243    secret_ref: BTreeMap<String, PbSecretRef>,
244}
245
246impl std::ops::Deref for WithOptionsSecResolved {
247    type Target = BTreeMap<String, String>;
248
249    fn deref(&self) -> &Self::Target {
250        &self.inner
251    }
252}
253
254impl std::ops::DerefMut for WithOptionsSecResolved {
255    fn deref_mut(&mut self) -> &mut Self::Target {
256        &mut self.inner
257    }
258}
259
260impl WithOptionsSecResolved {
261    /// Create a new [`WithOptions`] from a option [`BTreeMap`] and resolved secret ref.
262    pub fn new(inner: BTreeMap<String, String>, secret_ref: BTreeMap<String, PbSecretRef>) -> Self {
263        Self { inner, secret_ref }
264    }
265
266    pub fn as_plaintext(&self) -> &BTreeMap<String, String> {
267        &self.inner
268    }
269
270    pub fn as_secret(&self) -> &BTreeMap<String, PbSecretRef> {
271        &self.secret_ref
272    }
273
274    pub fn handle_update(
275        &mut self,
276        update_alter_props: BTreeMap<String, String>,
277        update_alter_secret_refs: BTreeMap<String, PbSecretRef>,
278    ) -> ConnectorResult<(Vec<SecretId>, Vec<SecretId>)> {
279        let old_secret_ids = self
280            .secret_ref
281            .values()
282            .map(|secret_ref| secret_ref.secret_id)
283            .collect::<BTreeSet<_>>();
284
285        // make sure the key in update_alter_props and update_alter_secret_refs not collide
286        for key in update_alter_props.keys() {
287            if update_alter_secret_refs.contains_key(key) {
288                return Err(
289                    anyhow::anyhow!("the key {} is set both in plaintext and secret", key).into(),
290                );
291            }
292        }
293
294        // remove legacy key if it's set in both plaintext and secret
295        // When a property changes from secret to plaintext, remove the old secret dependency
296        for k in update_alter_props.keys() {
297            self.secret_ref.remove(k);
298        }
299
300        // Handle secret ref updates
301        for k in update_alter_secret_refs.keys() {
302            // Remove any plaintext value for this key
303            self.inner.remove(k);
304        }
305
306        self.inner.extend(update_alter_props);
307        self.secret_ref.extend(update_alter_secret_refs);
308
309        let new_secret_ids = self
310            .secret_ref
311            .values()
312            .map(|secret_ref| secret_ref.secret_id)
313            .collect::<BTreeSet<_>>();
314        let to_add_secret_dep = new_secret_ids
315            .difference(&old_secret_ids)
316            .copied()
317            .collect();
318        let to_remove_secret_dep = old_secret_ids
319            .difference(&new_secret_ids)
320            .copied()
321            .collect();
322
323        Ok((to_add_secret_dep, to_remove_secret_dep))
324    }
325
326    /// Create a new [`WithOptions`] from a [`BTreeMap`].
327    pub fn without_secrets(inner: BTreeMap<String, String>) -> Self {
328        Self {
329            inner,
330            secret_ref: Default::default(),
331        }
332    }
333
334    /// Take the value of the option map and secret refs.
335    pub fn into_parts(self) -> (BTreeMap<String, String>, BTreeMap<String, PbSecretRef>) {
336        (self.inner, self.secret_ref)
337    }
338
339    pub fn value_eq_ignore_case(&self, key: &str, val: &str) -> bool {
340        if let Some(inner_val) = self.inner.get(key)
341            && inner_val.eq_ignore_ascii_case(val)
342        {
343            return true;
344        }
345        false
346    }
347}
348
349/// For `planner_test` crate so that it does not depend directly on `connector` crate just for `SinkFormatDesc`.
350impl TryFrom<&WithOptionsSecResolved> for Option<SinkFormatDesc> {
351    type Error = crate::sink::SinkError;
352
353    fn try_from(value: &WithOptionsSecResolved) -> std::result::Result<Self, Self::Error> {
354        let connector = value.get(crate::sink::CONNECTOR_TYPE_KEY);
355        let r#type = value.get(crate::sink::SINK_TYPE_OPTION);
356        match (connector, r#type) {
357            (Some(c), Some(t)) => SinkFormatDesc::from_legacy_type(c, t),
358            _ => Ok(None),
359        }
360    }
361}
362
363impl Get for WithOptionsSecResolved {
364    fn get(&self, key: &str) -> Option<&String> {
365        self.inner.get(key)
366    }
367}
368
369impl GetKeyIter for WithOptionsSecResolved {
370    fn key_iter(&self) -> impl Iterator<Item = &str> {
371        self.inner.keys().map(|s| s.as_str())
372    }
373}
374
375#[cfg(test)]
376mod tests {
377    use super::*;
378
379    fn source_options(connector: &str) -> WithOptionsSecResolved {
380        WithOptionsSecResolved::without_secrets(BTreeMap::from([(
381            UPSTREAM_SOURCE_KEY.to_owned(),
382            connector.to_owned(),
383        )]))
384    }
385
386    #[test]
387    fn test_full_reload_refresh_connector_whitelist() {
388        for connector in [
389            OPENDAL_S3_CONNECTOR,
390            GCS_CONNECTOR,
391            BATCH_POSIX_FS_CONNECTOR,
392            ICEBERG_CONNECTOR,
393            ADBC_SNOWFLAKE_CONNECTOR,
394        ] {
395            assert!(
396                source_options(connector).supports_full_reload_refresh(),
397                "{connector} should support FULL_RELOAD refresh"
398            );
399        }
400
401        for connector in [KAFKA_CONNECTOR, POSIX_FS_CONNECTOR, AZBLOB_CONNECTOR] {
402            assert!(
403                !source_options(connector).supports_full_reload_refresh(),
404                "{connector} should not support FULL_RELOAD refresh"
405            );
406        }
407    }
408}
risingwave_connector/with_options.rs

risingwave_connector/
with_options.rs
