Compare commits
1 Commits
e535bcc295
...
status_cod
| Author | SHA1 | Date | |
|---|---|---|---|
|
f7a3ca8fd7 |
1
.gitignore
vendored
1
.gitignore
vendored
@@ -6,4 +6,3 @@ flamegraph.svg
|
||||
perf.data.old
|
||||
/docker/logs/*
|
||||
/downloaded
|
||||
/Crawler.toml
|
||||
|
||||
429
Cargo.lock
generated
429
Cargo.lock
generated
@@ -103,18 +103,62 @@ dependencies = [
|
||||
"libc",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "anstream"
|
||||
version = "0.6.18"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "8acc5369981196006228e28809f761875c0327210a891e941f4c683b3a99529b"
|
||||
dependencies = [
|
||||
"anstyle",
|
||||
"anstyle-parse",
|
||||
"anstyle-query",
|
||||
"anstyle-wincon",
|
||||
"colorchoice",
|
||||
"is_terminal_polyfill",
|
||||
"utf8parse",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "anstyle"
|
||||
version = "1.0.10"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "55cc3b69f167a1ef2e161439aa98aed94e6028e5f9a59be9a6ffb47aef1651f9"
|
||||
|
||||
[[package]]
|
||||
name = "anstyle-parse"
|
||||
version = "0.2.6"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "3b2d16507662817a6a20a9ea92df6652ee4f94f914589377d69f3b21bc5798a9"
|
||||
dependencies = [
|
||||
"utf8parse",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "anstyle-query"
|
||||
version = "1.1.2"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "79947af37f4177cfead1110013d678905c37501914fba0efea834c3fe9a8d60c"
|
||||
dependencies = [
|
||||
"windows-sys 0.59.0",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "anstyle-wincon"
|
||||
version = "3.0.7"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "ca3534e77181a9cc07539ad51f2141fe32f6c3ffd4df76db8ad92346b003ae4e"
|
||||
dependencies = [
|
||||
"anstyle",
|
||||
"once_cell",
|
||||
"windows-sys 0.59.0",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "any_ascii"
|
||||
version = "0.3.2"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "ea50b14b7a4b9343f8c627a7a53c52076482bd4bdad0a24fd3ec533ed616cc2c"
|
||||
|
||||
[[package]]
|
||||
name = "anyhow"
|
||||
version = "1.0.98"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "e16d2d3311acee920a9eb8d33b8cbc1787ce4a264e85f964c2404b969bdcd487"
|
||||
|
||||
[[package]]
|
||||
name = "approx"
|
||||
version = "0.4.0"
|
||||
@@ -281,6 +325,17 @@ dependencies = [
|
||||
"serde_json",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "async-recursion"
|
||||
version = "1.1.1"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "3b43422f69d8ff38f95f1b2bb76517c91589a924d1559a0e935d7c8ce0274c11"
|
||||
dependencies = [
|
||||
"proc-macro2",
|
||||
"quote",
|
||||
"syn 2.0.100",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "async-stream"
|
||||
version = "0.3.6"
|
||||
@@ -760,6 +815,12 @@ dependencies = [
|
||||
"cc",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "colorchoice"
|
||||
version = "1.0.3"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "5b63caa9aa9397e2d9480a9b13673856c78d8ac123288526c37d7839f2a86990"
|
||||
|
||||
[[package]]
|
||||
name = "concurrent-queue"
|
||||
version = "2.5.0"
|
||||
@@ -810,6 +871,21 @@ dependencies = [
|
||||
"libc",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "crc"
|
||||
version = "3.2.1"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "69e6e4d7b33a94f0991c26729976b10ebde1d34c3ee82408fb536164fa10d636"
|
||||
dependencies = [
|
||||
"crc-catalog",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "crc-catalog"
|
||||
version = "2.4.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "19d374276b40fb8bbdee95aef7c7fa6b5316ec764510eb64b8dd0e2ed0d7e7f5"
|
||||
|
||||
[[package]]
|
||||
name = "crc32fast"
|
||||
version = "1.4.2"
|
||||
@@ -908,6 +984,20 @@ dependencies = [
|
||||
"parking_lot_core",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "dashmap"
|
||||
version = "6.1.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "5041cc499144891f3790297212f32a74fb938e5136a14943f338ef9e0ae276cf"
|
||||
dependencies = [
|
||||
"cfg-if",
|
||||
"crossbeam-utils",
|
||||
"hashbrown 0.14.5",
|
||||
"lock_api",
|
||||
"once_cell",
|
||||
"parking_lot_core",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "data-encoding"
|
||||
version = "2.8.0"
|
||||
@@ -924,6 +1014,17 @@ dependencies = [
|
||||
"serde",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "derivative"
|
||||
version = "2.2.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "fcc3dd5e9e9c0b295d6e1e4d811fb6f157d5ffd784b8d202fc62eac8035a770b"
|
||||
dependencies = [
|
||||
"proc-macro2",
|
||||
"quote",
|
||||
"syn 1.0.109",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "deunicode"
|
||||
version = "1.6.1"
|
||||
@@ -1034,6 +1135,29 @@ version = "0.1.2"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "c34f04666d835ff5d62e058c3995147c06f42fe86ff053337632bca83e42702d"
|
||||
|
||||
[[package]]
|
||||
name = "env_filter"
|
||||
version = "0.1.3"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "186e05a59d4c50738528153b83b0b0194d3a29507dfec16eccd4b342903397d0"
|
||||
dependencies = [
|
||||
"log",
|
||||
"regex",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "env_logger"
|
||||
version = "0.11.7"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "c3716d7a920fb4fac5d84e9d4bce8ceb321e9414b4409da61b07b75c1e3d0697"
|
||||
dependencies = [
|
||||
"anstream",
|
||||
"anstyle",
|
||||
"env_filter",
|
||||
"jiff",
|
||||
"log",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "equivalent"
|
||||
version = "1.0.2"
|
||||
@@ -1603,19 +1727,6 @@ dependencies = [
|
||||
"webpki-roots",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "hyper-timeout"
|
||||
version = "0.5.2"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "2b90d566bffbce6a75bd8b09a05aa8c2cb1fabb6cb348f8840c9e4c90a0d83b0"
|
||||
dependencies = [
|
||||
"hyper",
|
||||
"hyper-util",
|
||||
"pin-project-lite",
|
||||
"tokio",
|
||||
"tower-service",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "hyper-tls"
|
||||
version = "0.6.0"
|
||||
@@ -1859,10 +1970,7 @@ dependencies = [
|
||||
"html5ever 0.29.1",
|
||||
"metrics",
|
||||
"metrics-exporter-prometheus",
|
||||
"opentelemetry",
|
||||
"opentelemetry-otlp",
|
||||
"opentelemetry_sdk",
|
||||
"rand 0.9.1",
|
||||
"minio",
|
||||
"reqwest",
|
||||
"serde",
|
||||
"surrealdb",
|
||||
@@ -1879,6 +1987,12 @@ version = "2.11.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "469fb0b9cefa57e3ef31275ee7cacb78f2fdca44e4765491884a2b119d4eb130"
|
||||
|
||||
[[package]]
|
||||
name = "is_terminal_polyfill"
|
||||
version = "1.70.1"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "7943c866cc5cd64cbc25b2e01621d07fa8eb2a1a23160ee81ce38704e97b8ecf"
|
||||
|
||||
[[package]]
|
||||
name = "itertools"
|
||||
version = "0.10.5"
|
||||
@@ -1912,6 +2026,30 @@ version = "1.0.15"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "4a5f13b858c8d314ee3e8f639011f7ccefe71f97f96e50151fb991f267928e2c"
|
||||
|
||||
[[package]]
|
||||
name = "jiff"
|
||||
version = "0.2.4"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "d699bc6dfc879fb1bf9bdff0d4c56f0884fc6f0d0eb0fba397a6d00cd9a6b85e"
|
||||
dependencies = [
|
||||
"jiff-static",
|
||||
"log",
|
||||
"portable-atomic",
|
||||
"portable-atomic-util",
|
||||
"serde",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "jiff-static"
|
||||
version = "0.2.4"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "8d16e75759ee0aa64c57a56acbf43916987b20c77373cb7e808979e02b93c9f9"
|
||||
dependencies = [
|
||||
"proc-macro2",
|
||||
"quote",
|
||||
"syn 2.0.100",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "jobserver"
|
||||
version = "0.1.32"
|
||||
@@ -2156,6 +2294,12 @@ dependencies = [
|
||||
"digest",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "md5"
|
||||
version = "0.7.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "490cc448043f947bae3cbee9c203358d62dbee0db12107a74be5c30ccfd09771"
|
||||
|
||||
[[package]]
|
||||
name = "memchr"
|
||||
version = "2.7.4"
|
||||
@@ -2254,6 +2398,46 @@ version = "0.2.1"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "68354c5c6bd36d73ff3feceb05efa59b6acb7626617f4962be322a825e61f79a"
|
||||
|
||||
[[package]]
|
||||
name = "minio"
|
||||
version = "0.2.0-alpha"
|
||||
source = "git+https://github.com/minio/minio-rs.git?rev=c28f576#c28f576cb8f8cf47fb941bb9db62b2cbd6f080c1"
|
||||
dependencies = [
|
||||
"async-recursion",
|
||||
"async-trait",
|
||||
"base64 0.22.1",
|
||||
"byteorder",
|
||||
"bytes",
|
||||
"chrono",
|
||||
"crc",
|
||||
"dashmap 6.1.0",
|
||||
"derivative",
|
||||
"env_logger",
|
||||
"futures-util",
|
||||
"hex",
|
||||
"hmac",
|
||||
"home",
|
||||
"http",
|
||||
"hyper",
|
||||
"lazy_static",
|
||||
"log",
|
||||
"md5",
|
||||
"multimap",
|
||||
"os_info",
|
||||
"percent-encoding",
|
||||
"rand 0.8.5",
|
||||
"regex",
|
||||
"reqwest",
|
||||
"serde",
|
||||
"serde_json",
|
||||
"sha2",
|
||||
"tokio",
|
||||
"tokio-stream",
|
||||
"tokio-util",
|
||||
"urlencoding",
|
||||
"xmltree",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "miniz_oxide"
|
||||
version = "0.8.5"
|
||||
@@ -2291,6 +2475,15 @@ dependencies = [
|
||||
"version_check",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "multimap"
|
||||
version = "0.10.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "defc4c55412d89136f966bbb339008b474350e5e6e78d2714439c386b3137a03"
|
||||
dependencies = [
|
||||
"serde",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "nanoid"
|
||||
version = "0.4.0"
|
||||
@@ -2543,77 +2736,14 @@ dependencies = [
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "opentelemetry"
|
||||
version = "0.30.0"
|
||||
name = "os_info"
|
||||
version = "3.10.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "aaf416e4cb72756655126f7dd7bb0af49c674f4c1b9903e80c009e0c37e552e6"
|
||||
checksum = "2a604e53c24761286860eba4e2c8b23a0161526476b1de520139d69cdb85a6b5"
|
||||
dependencies = [
|
||||
"futures-core",
|
||||
"futures-sink",
|
||||
"js-sys",
|
||||
"pin-project-lite",
|
||||
"thiserror 2.0.12",
|
||||
"tracing",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "opentelemetry-http"
|
||||
version = "0.30.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "50f6639e842a97dbea8886e3439710ae463120091e2e064518ba8e716e6ac36d"
|
||||
dependencies = [
|
||||
"async-trait",
|
||||
"bytes",
|
||||
"http",
|
||||
"opentelemetry",
|
||||
"reqwest",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "opentelemetry-otlp"
|
||||
version = "0.30.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "dbee664a43e07615731afc539ca60c6d9f1a9425e25ca09c57bc36c87c55852b"
|
||||
dependencies = [
|
||||
"http",
|
||||
"opentelemetry",
|
||||
"opentelemetry-http",
|
||||
"opentelemetry-proto",
|
||||
"opentelemetry_sdk",
|
||||
"prost",
|
||||
"reqwest",
|
||||
"thiserror 2.0.12",
|
||||
"tokio",
|
||||
"tonic",
|
||||
"tracing",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "opentelemetry-proto"
|
||||
version = "0.30.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "2e046fd7660710fe5a05e8748e70d9058dc15c94ba914e7c4faa7c728f0e8ddc"
|
||||
dependencies = [
|
||||
"opentelemetry",
|
||||
"opentelemetry_sdk",
|
||||
"prost",
|
||||
"tonic",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "opentelemetry_sdk"
|
||||
version = "0.30.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "11f644aa9e5e31d11896e024305d7e3c98a88884d9f8919dbf37a9991bc47a4b"
|
||||
dependencies = [
|
||||
"futures-channel",
|
||||
"futures-executor",
|
||||
"futures-util",
|
||||
"opentelemetry",
|
||||
"percent-encoding",
|
||||
"rand 0.9.1",
|
||||
"serde_json",
|
||||
"thiserror 2.0.12",
|
||||
"log",
|
||||
"serde",
|
||||
"windows-sys 0.52.0",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
@@ -2793,26 +2923,6 @@ version = "0.5.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "5be167a7af36ee22fe3115051bc51f6e6c7054c9348e28deb4f49bd6f705a315"
|
||||
|
||||
[[package]]
|
||||
name = "pin-project"
|
||||
version = "1.1.10"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "677f1add503faace112b9f1373e43e9e054bfdd22ff1a63c1bc485eaec6a6a8a"
|
||||
dependencies = [
|
||||
"pin-project-internal",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "pin-project-internal"
|
||||
version = "1.1.10"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "6e918e4ff8c4549eb882f14b3a4bc8c8bc93de829416eacf579f1207a8fbf861"
|
||||
dependencies = [
|
||||
"proc-macro2",
|
||||
"quote",
|
||||
"syn 2.0.100",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "pin-project-lite"
|
||||
version = "0.2.16"
|
||||
@@ -2837,6 +2947,15 @@ version = "1.11.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "350e9b48cbc6b0e028b0473b114454c6316e57336ee184ceab6e53f72c178b3e"
|
||||
|
||||
[[package]]
|
||||
name = "portable-atomic-util"
|
||||
version = "0.2.4"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "d8a2f0d8d040d7848a709caf78912debcc3f33ee4b3cac47d73d1e1069e83507"
|
||||
dependencies = [
|
||||
"portable-atomic",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "powerfmt"
|
||||
version = "0.2.0"
|
||||
@@ -2886,29 +3005,6 @@ dependencies = [
|
||||
"unicode-ident",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "prost"
|
||||
version = "0.13.5"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "2796faa41db3ec313a31f7624d9286acf277b52de526150b7e69f3debf891ee5"
|
||||
dependencies = [
|
||||
"bytes",
|
||||
"prost-derive",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "prost-derive"
|
||||
version = "0.13.5"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "8a56d757972c98b346a9b766e3f02746cde6dd1cd1d1d563472929fdd74bec4d"
|
||||
dependencies = [
|
||||
"anyhow",
|
||||
"itertools 0.13.0",
|
||||
"proc-macro2",
|
||||
"quote",
|
||||
"syn 2.0.100",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "psl-types"
|
||||
version = "2.0.11"
|
||||
@@ -2999,7 +3095,7 @@ checksum = "b820744eb4dc9b57a3398183639c511b5a26d2ed702cedd3febaa1393caa22cc"
|
||||
dependencies = [
|
||||
"bytes",
|
||||
"getrandom 0.3.2",
|
||||
"rand 0.9.1",
|
||||
"rand 0.9.0",
|
||||
"ring",
|
||||
"rustc-hash 2.1.1",
|
||||
"rustls",
|
||||
@@ -3070,12 +3166,13 @@ dependencies = [
|
||||
|
||||
[[package]]
|
||||
name = "rand"
|
||||
version = "0.9.1"
|
||||
version = "0.9.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "9fbfd9d094a40bf3ae768db9361049ace4c0e04a4fd6b359518bd7b73a73dd97"
|
||||
checksum = "3779b94aeb87e8bd4e834cee3650289ee9e0d5677f976ecdb6d219e5f4f6cd94"
|
||||
dependencies = [
|
||||
"rand_chacha 0.9.0",
|
||||
"rand_core 0.9.3",
|
||||
"zerocopy 0.8.23",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
@@ -3269,7 +3366,6 @@ dependencies = [
|
||||
"base64 0.22.1",
|
||||
"bytes",
|
||||
"encoding_rs",
|
||||
"futures-channel",
|
||||
"futures-core",
|
||||
"futures-util",
|
||||
"h2",
|
||||
@@ -4100,7 +4196,7 @@ dependencies = [
|
||||
"cedar-policy",
|
||||
"chrono",
|
||||
"ciborium",
|
||||
"dashmap",
|
||||
"dashmap 5.5.3",
|
||||
"deunicode",
|
||||
"dmp",
|
||||
"fst",
|
||||
@@ -4523,32 +4619,6 @@ dependencies = [
|
||||
"winnow",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "tonic"
|
||||
version = "0.13.1"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "7e581ba15a835f4d9ea06c55ab1bd4dce26fc53752c69a04aac00703bfb49ba9"
|
||||
dependencies = [
|
||||
"async-trait",
|
||||
"base64 0.22.1",
|
||||
"bytes",
|
||||
"http",
|
||||
"http-body",
|
||||
"http-body-util",
|
||||
"hyper",
|
||||
"hyper-timeout",
|
||||
"hyper-util",
|
||||
"percent-encoding",
|
||||
"pin-project",
|
||||
"prost",
|
||||
"tokio",
|
||||
"tokio-stream",
|
||||
"tower",
|
||||
"tower-layer",
|
||||
"tower-service",
|
||||
"tracing",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "tower"
|
||||
version = "0.5.2"
|
||||
@@ -4557,15 +4627,11 @@ checksum = "d039ad9159c98b70ecfd540b2573b97f7f52c3e8d9f8ad57a24b916a536975f9"
|
||||
dependencies = [
|
||||
"futures-core",
|
||||
"futures-util",
|
||||
"indexmap 2.8.0",
|
||||
"pin-project-lite",
|
||||
"slab",
|
||||
"sync_wrapper",
|
||||
"tokio",
|
||||
"tokio-util",
|
||||
"tower-layer",
|
||||
"tower-service",
|
||||
"tracing",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
@@ -4711,7 +4777,7 @@ version = "1.2.1"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "470dbf6591da1b39d43c14523b2b469c86879a53e8b758c8e090a470fe7b1fbe"
|
||||
dependencies = [
|
||||
"rand 0.9.1",
|
||||
"rand 0.9.0",
|
||||
"serde",
|
||||
"web-time",
|
||||
]
|
||||
@@ -4807,6 +4873,12 @@ version = "1.0.4"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "b6c140620e7ffbb22c2dee59cafe6084a59b5ffc27a8859a5f0d494b5d52b6be"
|
||||
|
||||
[[package]]
|
||||
name = "utf8parse"
|
||||
version = "0.2.2"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "06abde3611657adf66d383f00b093d7faecc7fa57071cce2578660c9f1010821"
|
||||
|
||||
[[package]]
|
||||
name = "uuid"
|
||||
version = "1.16.0"
|
||||
@@ -5347,6 +5419,21 @@ dependencies = [
|
||||
"tap",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "xml-rs"
|
||||
version = "0.8.25"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "c5b940ebc25896e71dd073bad2dbaa2abfe97b0a391415e22ad1326d9c54e3c4"
|
||||
|
||||
[[package]]
|
||||
name = "xmltree"
|
||||
version = "0.11.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "b619f8c85654798007fb10afa5125590b43b088c225a25fc2fec100a9fad0fc6"
|
||||
dependencies = [
|
||||
"xml-rs",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "yoke"
|
||||
version = "0.7.5"
|
||||
|
||||
@@ -9,10 +9,8 @@ futures-util = "0.3.31"
|
||||
html5ever = "0.29"
|
||||
metrics = "0.24.1"
|
||||
metrics-exporter-prometheus = { version = "0.16.2", features=["http-listener"]}
|
||||
opentelemetry = "0.30.0"
|
||||
opentelemetry-otlp = { version = "0.30.0", features = ["metrics", "trace", "logs", "grpc-tonic"] }
|
||||
opentelemetry_sdk = "0.30.0"
|
||||
rand = "0.9.1"
|
||||
# minio = "0.1.0"
|
||||
minio = {git="https://github.com/minio/minio-rs.git", rev = "c28f576"}
|
||||
reqwest = { version = "0.12", features = ["gzip", "default", "rustls-tls", "stream"] }
|
||||
serde = { version = "1.0", features = ["derive"] }
|
||||
surrealdb = "2.2"
|
||||
|
||||
18
Crawler.toml
18
Crawler.toml
@@ -1,11 +1,3 @@
|
||||
# Visability config
|
||||
# Alloy (for Tempo)
|
||||
tracing_endpoint = "http://localhost:4317"
|
||||
# Prometheus
|
||||
metrics_endpoint = "http://localhost:9090/api/v1/otlp/v1/metrics"
|
||||
# Alloy (for Loki)
|
||||
log_file = "./docker/logs/tracing.log"
|
||||
|
||||
# Surreal config
|
||||
surreal_url = "localhost:8000"
|
||||
surreal_username = "root"
|
||||
@@ -14,9 +6,7 @@ surreal_ns = "test"
|
||||
surreal_db = "v1.21.1"
|
||||
|
||||
# Crawler config
|
||||
crawl_filter = "https://ftpgeoinfo.msl.mt.gov/Data/Spatial/MSDI/Imagery/2023_NAIP/UTM_County_Mosaics/"
|
||||
# crawl_filter = "https://oliveratkinson.net"
|
||||
start_url = "https://ftpgeoinfo.msl.mt.gov/Data/Spatial/MSDI/Imagery/2023_NAIP/UTM_County_Mosaics/"
|
||||
# start_url = "https://oliveratkinson.net"
|
||||
budget = 100
|
||||
batch_size = 2
|
||||
crawl_filter = "https://ftpgeoinfo.msl.mt.gov/Data/Spatial/MSDI"
|
||||
start_url = "https://ftpgeoinfo.msl.mt.gov/Data/Spatial/MSDI"
|
||||
budget = 10000
|
||||
batch_size = 50
|
||||
|
||||
@@ -12,25 +12,3 @@ loki.write "local_loki" {
|
||||
url = "http://loki:3100/loki/api/v1/push"
|
||||
}
|
||||
}
|
||||
|
||||
otelcol.receiver.otlp "otlp_receiver" {
|
||||
grpc {
|
||||
endpoint = "0.0.0.0:4317"
|
||||
}
|
||||
http {
|
||||
endpoint = "0.0.0.0:4318"
|
||||
}
|
||||
|
||||
output {
|
||||
traces = [otelcol.exporter.otlp.tempo.input,]
|
||||
}
|
||||
}
|
||||
|
||||
otelcol.exporter.otlp "tempo" {
|
||||
client {
|
||||
endpoint = "tempo:4317"
|
||||
tls {
|
||||
insecure = true
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1,6 +1,4 @@
|
||||
services:
|
||||
|
||||
# Database
|
||||
surreal:
|
||||
image: surrealdb/surrealdb:latest-dev
|
||||
ports:
|
||||
@@ -17,18 +15,6 @@ services:
|
||||
- root
|
||||
- rocksdb:/mydata/database.db
|
||||
|
||||
# Tracing
|
||||
tempo:
|
||||
image: grafana/tempo:latest
|
||||
command: [ "-config.file=/etc/tempo.yaml" ]
|
||||
volumes:
|
||||
- ./tempo.yaml:/etc/tempo.yaml
|
||||
- tempo_storage:/var/tempo
|
||||
ports:
|
||||
- 3200:3200 # self metrics for prometheus
|
||||
- 4317:4317 # otlp grpc - (alloy)
|
||||
|
||||
# Log scraper
|
||||
alloy:
|
||||
image: grafana/alloy:latest
|
||||
ports:
|
||||
@@ -38,13 +24,9 @@ services:
|
||||
- ./logs/:/tmp/alloy-logs
|
||||
- ./alloy.conf:/etc/alloy/config.alloy
|
||||
- alloy_storage:/var/lib/alloy
|
||||
command:
|
||||
- run
|
||||
- --server.http.listen-addr=0.0.0.0:12345
|
||||
- --storage.path=/var/lib/alloy/data
|
||||
- /etc/alloy/config.alloy
|
||||
command: run --server.http.listen-addr=0.0.0.0:12345 --storage.path=/var/lib/alloy/data /etc/alloy/config.alloy
|
||||
|
||||
# Log storage / analysis
|
||||
#logs
|
||||
loki:
|
||||
image: grafana/loki:latest
|
||||
ports:
|
||||
@@ -53,21 +35,16 @@ services:
|
||||
volumes:
|
||||
- ./loki.yaml:/etc/loki/local-config.yaml
|
||||
|
||||
# Metrics
|
||||
# Metrics collector
|
||||
prometheus:
|
||||
image: prom/prometheus:latest
|
||||
ports:
|
||||
- 9090:9090
|
||||
expose:
|
||||
- 9090
|
||||
volumes:
|
||||
- ./prometheus.yaml:/etc/prometheus/prometheus.yml
|
||||
# persist data
|
||||
# - prometheus_storage:/prometheus
|
||||
command:
|
||||
- --enable-feature=native-histograms
|
||||
- --web.enable-remote-write-receiver
|
||||
- --web.enable-lifecycle
|
||||
- --web.enable-otlp-receiver
|
||||
- --config.file=/etc/prometheus/prometheus.yml
|
||||
- prometheus_storage:/prometheus
|
||||
command: --web.enable-lifecycle --config.file=/etc/prometheus/prometheus.yml
|
||||
|
||||
# Everything viewer
|
||||
grafana:
|
||||
@@ -89,4 +66,3 @@ volumes:
|
||||
grafana_storage:
|
||||
alloy_storage:
|
||||
surrealdb_storage:
|
||||
tempo_storage:
|
||||
|
||||
@@ -22,20 +22,3 @@ datasources:
|
||||
editable: false
|
||||
jsonData:
|
||||
httpMethod: GET
|
||||
- name: Tempo
|
||||
type: tempo
|
||||
access: proxy
|
||||
orgId: 1
|
||||
url: http://tempo:3200
|
||||
basicAuth: false
|
||||
isDefault: false
|
||||
version: 1
|
||||
editable: true
|
||||
apiVersion: 1
|
||||
uid: tempo
|
||||
jsonData:
|
||||
httpMethod: GET
|
||||
serviceMap:
|
||||
datasourceUid: prometheus
|
||||
streamingEnabled:
|
||||
search: true
|
||||
|
||||
@@ -1,15 +1,17 @@
|
||||
global:
|
||||
scrape_interval: 60s
|
||||
scrape_interval: 5s
|
||||
query_log_file: /etc/prometheus/query.log
|
||||
|
||||
scrape_configs:
|
||||
# Crawler configs get pushed with OTLP
|
||||
# - job_name: 'loki'
|
||||
# static_configs:
|
||||
# - targets: ['loki:3100']
|
||||
# - job_name: 'prometheus'
|
||||
# static_configs:
|
||||
# - targets: ['localhost:9090']
|
||||
- job_name: 'tempo'
|
||||
- job_name: crawler
|
||||
static_configs:
|
||||
- targets: ['tempo:3200']
|
||||
# change this your machine's ip, localhost won't work
|
||||
# because localhost refers to the docker container.
|
||||
- targets: ['192.168.1.200:2500']
|
||||
#- targets: ['192.168.8.209:2500']
|
||||
- job_name: loki
|
||||
static_configs:
|
||||
- targets: ['loki:3100']
|
||||
- job_name: prometheus
|
||||
static_configs:
|
||||
- targets: ['localhost:9090']
|
||||
|
||||
@@ -1,48 +0,0 @@
|
||||
stream_over_http_enabled: true
|
||||
server:
|
||||
http_listen_port: 3200
|
||||
log_level: info
|
||||
|
||||
query_frontend:
|
||||
search:
|
||||
duration_slo: 5s
|
||||
throughput_bytes_slo: 1.073741824e+09
|
||||
metadata_slo:
|
||||
duration_slo: 5s
|
||||
throughput_bytes_slo: 1.073741824e+09
|
||||
trace_by_id:
|
||||
duration_slo: 5s
|
||||
|
||||
distributor:
|
||||
receivers:
|
||||
otlp:
|
||||
protocols:
|
||||
grpc:
|
||||
endpoint: "tempo:4317"
|
||||
|
||||
metrics_generator:
|
||||
registry:
|
||||
external_labels:
|
||||
source: tempo
|
||||
cluster: docker-compose
|
||||
storage:
|
||||
path: /var/tempo/generator/wal
|
||||
remote_write:
|
||||
- url: http://prometheus:9090/api/v1/write
|
||||
send_exemplars: true
|
||||
traces_storage:
|
||||
path: /var/tempo/generator/traces
|
||||
|
||||
storage:
|
||||
trace:
|
||||
backend: local # backend configuration to use
|
||||
wal:
|
||||
path: /var/tempo/wal # where to store the wal locally
|
||||
local:
|
||||
path: /var/tempo/blocks
|
||||
|
||||
overrides:
|
||||
defaults:
|
||||
metrics_generator:
|
||||
processors: [service-graphs, span-metrics, local-blocks] # enables metrics generator
|
||||
generate_native_histograms: both
|
||||
50
src/db.rs
50
src/db.rs
@@ -1,13 +1,12 @@
|
||||
use metrics::counter;
|
||||
use std::fmt::Debug;
|
||||
use serde::{Deserialize, Serialize};
|
||||
use std::{fmt::Debug, time::Duration};
|
||||
use surrealdb::{
|
||||
engine::remote::ws::{Client, Ws},
|
||||
opt::auth::Root,
|
||||
sql::Thing,
|
||||
Surreal,
|
||||
};
|
||||
use tokio::time::sleep;
|
||||
use tracing::{error, instrument, trace};
|
||||
use url::Url;
|
||||
|
||||
@@ -17,7 +16,6 @@ const STORE: &str = "surql_store_calls";
|
||||
|
||||
#[derive(Serialize, Deserialize, Clone, Eq, PartialEq, Hash)]
|
||||
pub struct Website {
|
||||
pub id: Option<Thing>,
|
||||
/// The url that this data is found at
|
||||
pub site: Url,
|
||||
/// Wether or not this link has been crawled yet
|
||||
@@ -48,7 +46,6 @@ impl Website {
|
||||
crawled,
|
||||
site,
|
||||
status_code: 0,
|
||||
id: None,
|
||||
}
|
||||
}
|
||||
|
||||
@@ -59,14 +56,13 @@ impl Website {
|
||||
counter!(STORE).increment(1);
|
||||
let mut things = Vec::with_capacity(all.len());
|
||||
|
||||
// FIXME fails *sometimes* because "Resource Busy"
|
||||
// FIXME failes *sometimes* because "Resource Busy"
|
||||
match db
|
||||
.query(
|
||||
"INSERT INTO website $array
|
||||
ON DUPLICATE KEY UPDATE
|
||||
accessed_at = time::now(),
|
||||
status_code = $input.status_code,
|
||||
processing = false,
|
||||
crawled = crawled OR $input.crawled
|
||||
RETURN VALUE id;
|
||||
",
|
||||
@@ -86,47 +82,18 @@ impl Website {
|
||||
}
|
||||
}
|
||||
|
||||
/// Returns uncrawled links
|
||||
#[instrument(skip(db, config))]
|
||||
pub async fn get_next(db: &Surreal<Client>, config: &Config) -> Option<Website> {
|
||||
let mut res: Option<Website> = None;
|
||||
let mut fails = 0;
|
||||
|
||||
while res == None {
|
||||
let mut response = db
|
||||
.query("fn::get_next($format)")
|
||||
.bind(("format", config.crawl_filter.to_string()))
|
||||
.await
|
||||
.expect("Hard-coded query failed..?");
|
||||
|
||||
res = match response.take(0) {
|
||||
Ok(ok) => ok,
|
||||
Err(_err) => {
|
||||
// basically just CSMA/CA
|
||||
let delay = rand::random_range(10..500);
|
||||
sleep(Duration::from_millis(delay)).await;
|
||||
fails += 1;
|
||||
// Don't get stuck here forever, failing...
|
||||
// (most I've seen is 1)
|
||||
if fails > 5 {
|
||||
error!("Max attempts to get_next() reached... ({fails})");
|
||||
return None
|
||||
}
|
||||
None
|
||||
}
|
||||
};
|
||||
}
|
||||
|
||||
res
|
||||
}
|
||||
|
||||
#[derive(Debug, Serialize)]
|
||||
#[allow(dead_code)]
|
||||
pub struct Email {
|
||||
pub email: String,
|
||||
pub on: String,
|
||||
}
|
||||
|
||||
#[derive(Debug, Deserialize)]
|
||||
pub struct Record {
|
||||
#[allow(dead_code)]
|
||||
pub id: Thing,
|
||||
}
|
||||
|
||||
#[instrument(skip_all, name = "SurrealDB")]
|
||||
pub async fn connect(config: &Config) -> surrealdb::Result<Surreal<Client>> {
|
||||
trace!("Establishing connection to surreal...");
|
||||
@@ -155,3 +122,4 @@ pub async fn connect(config: &Config) -> surrealdb::Result<Surreal<Client>> {
|
||||
|
||||
Ok(db)
|
||||
}
|
||||
|
||||
|
||||
@@ -18,7 +18,7 @@ pub fn as_path(url: &Url, content_type: &HeaderValue) -> PathBuf {
|
||||
};
|
||||
|
||||
if let Some((ttype, subtype)) = ttype.split_once('/') {
|
||||
trace!(url = url.to_string(), main_type = ttype, sub_type = subtype, "Found Content-Type to be: {ttype}/{subtype}");
|
||||
trace!("Found Content-Type to be: {ttype}/{subtype} for {}", url.to_string());
|
||||
// If the Content-Type header is "*/html" (most likely "text/html") and the path's
|
||||
// extension is anything but html:
|
||||
if subtype=="html" && !url_path.extension().is_some_and(|f| f=="html" || f=="htm" ) {
|
||||
@@ -29,7 +29,7 @@ pub fn as_path(url: &Url, content_type: &HeaderValue) -> PathBuf {
|
||||
} else {
|
||||
warn!("Header: {:?} couldn't be parsed into a string!", content_type);
|
||||
}
|
||||
trace!(url = url.to_string(), path = &*url_path.to_string_lossy(), "Converted URL into path");
|
||||
trace!("Final path for {} is: {:?}", url, url_path);
|
||||
|
||||
url_path
|
||||
}
|
||||
@@ -49,6 +49,7 @@ pub async fn init(filename: &PathBuf) -> Option<fs::File> {
|
||||
// create the folders
|
||||
if let Err(err) = fs::create_dir_all(&parent).await {
|
||||
error!("Dir creation: {err} {:?}", filename);
|
||||
eprintln!("{}", err)
|
||||
} else if let Ok(ok) = file().await {
|
||||
return Some(ok);
|
||||
}
|
||||
|
||||
510
src/main.rs
510
src/main.rs
@@ -1,79 +1,38 @@
|
||||
#![feature(ip_from)]
|
||||
#![feature(path_add_extension)]
|
||||
#![warn(clippy::expect_used)]
|
||||
#![deny(clippy::unwrap_used)]
|
||||
|
||||
extern crate html5ever;
|
||||
|
||||
use futures_util::StreamExt;
|
||||
|
||||
use std::{
|
||||
collections::HashSet, fs::File, io::Read, sync::{Arc, LazyLock}
|
||||
collections::HashSet,
|
||||
fs::File,
|
||||
io::Read,
|
||||
net::{IpAddr, Ipv4Addr},
|
||||
};
|
||||
|
||||
use futures_util::StreamExt;
|
||||
use opentelemetry::{global::{self}, metrics::{Counter, Meter, UpDownCounter}};
|
||||
use opentelemetry_otlp::{Protocol, WithExportConfig};
|
||||
use db::{connect, Website};
|
||||
use metrics::{counter, gauge};
|
||||
use metrics_exporter_prometheus::PrometheusBuilder;
|
||||
use serde::Deserialize;
|
||||
use surrealdb::{engine::remote::ws::Client, Surreal};
|
||||
use tokio::{io::{AsyncWriteExt, BufWriter}, sync::RwLock, task::JoinSet};
|
||||
use tracing::{debug, error, info, instrument, level_filters::LevelFilter, trace, warn};
|
||||
use tokio::{io::{AsyncWriteExt, BufWriter}, task::JoinSet};
|
||||
use tracing::{debug, debug_span, error, info, instrument, level_filters::LevelFilter, trace, trace_span, warn};
|
||||
use tracing_subscriber::{fmt, layer::SubscriberExt, EnvFilter, Layer, Registry};
|
||||
|
||||
use crate::db::get_next;
|
||||
|
||||
mod db;
|
||||
mod filesystem;
|
||||
mod parser;
|
||||
|
||||
static METER: LazyLock<Meter> = LazyLock::new(|| global::meter("Internet_Mapper"));
|
||||
static BEING_PROCESSED: LazyLock<UpDownCounter<i64>> = LazyLock::new(||
|
||||
METER
|
||||
.i64_up_down_counter("crawler_pages_being_processed")
|
||||
.build()
|
||||
);
|
||||
static BEING_PARSED: LazyLock<UpDownCounter<i64>> = LazyLock::new(||
|
||||
METER
|
||||
.i64_up_down_counter("crawler_pages_being_parsed")
|
||||
.build()
|
||||
);
|
||||
static BEING_STREAMED: LazyLock<UpDownCounter<i64>> = LazyLock::new(||
|
||||
METER
|
||||
.i64_up_down_counter("crawler_pages_being_streamed")
|
||||
.build()
|
||||
);
|
||||
static GET_IN_FLIGHT: LazyLock<UpDownCounter<i64>> = LazyLock::new(||
|
||||
METER
|
||||
.i64_up_down_counter("crawler_gets_in_flight")
|
||||
.build()
|
||||
);
|
||||
static TOTAL_BYTES_DOWN: LazyLock<Counter<u64>> = LazyLock::new(||
|
||||
METER
|
||||
.u64_counter("crawler_total_bytes_down")
|
||||
.build()
|
||||
);
|
||||
static SITES_CRAWLED: LazyLock<Counter<u64>> = LazyLock::new(||
|
||||
METER
|
||||
.u64_counter("crawler_total_sites_crawled")
|
||||
.build()
|
||||
);
|
||||
|
||||
static CONFIG: LazyLock<Config> = LazyLock::new(|| {
|
||||
let mut file = File::open("./Crawler.toml").expect("Failed to read Crawler.toml");
|
||||
let mut buf = String::new();
|
||||
let _ = file.read_to_string(&mut buf);
|
||||
|
||||
let config: Config = toml::from_str(&buf).expect("Failed to parse Crawler.toml");
|
||||
config
|
||||
});
|
||||
|
||||
// FIXME Traces aren't working on multiple threads, they block
|
||||
// static TRACER: LazyLock<BoxedTracer> = LazyLock::new(|| global::tracer("Internet_Mapper"));
|
||||
const GET_METRIC: &str = "total_gets";
|
||||
const GET_IN_FLIGHT: &str = "gets_in_flight";
|
||||
const SITES_CRAWLED: &str = "pages_crawled";
|
||||
const BEING_PROCESSED: &str = "pages_being_processed";
|
||||
|
||||
#[derive(Deserialize)]
|
||||
struct Config {
|
||||
tracing_endpoint: String,
|
||||
metrics_endpoint: String,
|
||||
log_file: String,
|
||||
|
||||
surreal_ns: String,
|
||||
surreal_db: String,
|
||||
surreal_url: String,
|
||||
@@ -90,224 +49,10 @@ struct Config {
|
||||
async fn main() {
|
||||
println!("Logs and metrics are provided to the Grafana dashboard");
|
||||
|
||||
// Start TRACE / LOGGING / METRICS
|
||||
load_tracing(&CONFIG);
|
||||
load_logging(&CONFIG);
|
||||
load_metrics(&CONFIG);
|
||||
|
||||
// When getting uncrawled pages, name must contain this variable. "" will effectively get ignored.
|
||||
// let crawl_filter = "en.wikipedia.org/";
|
||||
// let budget = 50;
|
||||
let crawled = Arc::new(RwLock::new(0));
|
||||
|
||||
let starting_url = &CONFIG.start_url;
|
||||
|
||||
let db = connect(&CONFIG)
|
||||
.await
|
||||
.expect("Failed to connect to surreal, aborting.");
|
||||
|
||||
let reqwest = reqwest::Client::builder()
|
||||
// .use_rustls_tls()
|
||||
.gzip(true)
|
||||
.build()
|
||||
.expect("Failed to build reqwest client.");
|
||||
|
||||
// Kick off the whole machine - This Website object doesn't matter, it's just to allow for
|
||||
// get() to work.
|
||||
// let mut span = TRACER.start("Pre-Loop");
|
||||
let site = Website::new(starting_url, false);
|
||||
process(site, db.clone(), reqwest.clone()).await;
|
||||
// span.end();
|
||||
|
||||
// let mut main_loop_span= TRACER.start("Main-Loop");
|
||||
let mut futures = JoinSet::new();
|
||||
for _ in 0..CONFIG.batch_size {
|
||||
futures.spawn(process_single_thread(&CONFIG, db.clone(), reqwest.clone(), crawled.clone()));
|
||||
}
|
||||
futures.join_all().await;
|
||||
// main_loop_span.end();
|
||||
|
||||
info!("Done");
|
||||
}
|
||||
|
||||
async fn process_single_thread(config: &Config, db: Surreal<Client>, reqwest: reqwest::Client, crawled: Arc<RwLock<usize>>) {
|
||||
while *(crawled.read().await) < config.budget {
|
||||
let uncrawled = get_next(&db.clone(), &config).await;
|
||||
match uncrawled {
|
||||
Some(site) => {
|
||||
process(site, db.clone(), reqwest.clone()).await;
|
||||
SITES_CRAWLED.add(1, &[]);
|
||||
// Somehow this write doesn't hang on the while's read?
|
||||
let mut c = crawled.write().await;
|
||||
*c += 1;
|
||||
},
|
||||
None => {
|
||||
warn!("fn::get_next() returned None");
|
||||
return;
|
||||
},
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
#[instrument(skip(db, reqwest))]
|
||||
/// Downloads and crawls and stores a webpage.
|
||||
/// It is acceptable to clone `db`, `reqwest`, and `s3` because they all use `Arc`s internally. - Noted by Oliver
|
||||
async fn process(mut site: Website, db: Surreal<Client>, reqwest: reqwest::Client) {
|
||||
|
||||
// METRICS
|
||||
debug!(url = &site.site.as_str(), "Process: {}", &site.site);
|
||||
BEING_PROCESSED.add(1, &[]);
|
||||
// let mut process_span = TRACER.start("Process");
|
||||
|
||||
// Build the request
|
||||
let request_builder = reqwest.get(site.site.to_string());
|
||||
|
||||
// Send the http request (get)
|
||||
GET_IN_FLIGHT.add(1, &[]);
|
||||
if let Ok(response) = request_builder.send().await {
|
||||
GET_IN_FLIGHT.add(-1, &[]);
|
||||
|
||||
let headers = response.headers();
|
||||
let code = response.status();
|
||||
if code != 200 {
|
||||
warn!("{code} for {}", site.site.as_str());
|
||||
}
|
||||
|
||||
#[allow(non_snake_case)]
|
||||
let CT = headers.get("Content-Type");
|
||||
let ct = headers.get("content-type");
|
||||
|
||||
let ct = match (CT,ct) {
|
||||
(None, None) => {
|
||||
warn!("Server did not respond with Content-Type header. Url: {} Headers: ({:?})", site.site.to_string(), headers);
|
||||
return
|
||||
},
|
||||
(None, Some(a)) => a,
|
||||
(Some(a), None) => a,
|
||||
(Some(a), Some(_)) => a,
|
||||
};
|
||||
|
||||
// create filepath (handles / -> /index.html)
|
||||
let path = filesystem::as_path(&site.site, ct);
|
||||
let mut tmp_path= path.clone();
|
||||
if !(tmp_path.add_extension("crawl_temp")) {
|
||||
warn!("Failed to add extension to file");
|
||||
// fallback ig
|
||||
tmp_path = tmp_path.with_extension("crawl_temp");
|
||||
}
|
||||
|
||||
// make sure that the file is good to go
|
||||
if let Some(file) = filesystem::init(&tmp_path).await {
|
||||
// Get body from response
|
||||
// stream the response onto the disk
|
||||
let mut stream = response.bytes_stream();
|
||||
|
||||
let should_parse = path.to_string_lossy().ends_with(".html");
|
||||
let mut writer = BufWriter::new(file);
|
||||
let mut buf: Vec<u8> = Vec::new();
|
||||
|
||||
// Write file to disk
|
||||
trace!("Writing at: {:?}", tmp_path);
|
||||
BEING_STREAMED.add(1, &[]);
|
||||
// let mut stream_span = TRACER.start("Stream");
|
||||
while let Some(data) = stream.next().await {
|
||||
match data {
|
||||
Ok(data) => {
|
||||
TOTAL_BYTES_DOWN.add(data.len() as u64, &[]);
|
||||
let _ = writer.write_all(&data).await;
|
||||
// If we are going to parse this file later, we will save it
|
||||
// into memory as well as the disk.
|
||||
// We do this because the data here might be incomplete
|
||||
if should_parse {
|
||||
data.iter().for_each(|f| buf.push(*f));
|
||||
}
|
||||
},
|
||||
Err(err) => {
|
||||
error!("{err}")
|
||||
},
|
||||
}
|
||||
}
|
||||
let _ = writer.flush().await;
|
||||
// rename the temp file into the real file name
|
||||
if let Err(err) = tokio::fs::rename(&tmp_path, &path).await {
|
||||
error!(
|
||||
from = &*tmp_path.to_string_lossy(),
|
||||
to = &*path.to_string_lossy(),
|
||||
"Error renaming file: {}",
|
||||
err
|
||||
);
|
||||
}
|
||||
|
||||
// stream_span.end();
|
||||
BEING_STREAMED.add(-1, &[]);
|
||||
|
||||
// (If needed) Parse the file
|
||||
if should_parse {
|
||||
BEING_PARSED.add(1, &[]);
|
||||
// let mut parsing_span = TRACER.start("Parsing");
|
||||
|
||||
// Parse document and get relationships
|
||||
let sites = parser::parse(&site, &buf).await;
|
||||
// De-duplicate this list
|
||||
let prev_len = sites.len();
|
||||
let set = sites.into_iter().fold(HashSet::new(), |mut set, item| {
|
||||
set.insert(item);
|
||||
set
|
||||
});
|
||||
let de_dupe_sites: Vec<Website> = set.into_iter().collect();
|
||||
let diff = prev_len - de_dupe_sites.len();
|
||||
trace!("Saved {diff} from being entered into the db by de-duping");
|
||||
// Store all the other sites so that we can link to them.
|
||||
let _ = Website::store_all(de_dupe_sites, &db).await;
|
||||
|
||||
// parsing_span.end();
|
||||
BEING_PARSED.add(-1, &[]);
|
||||
} else {
|
||||
trace!("Did not parse: {}", site.site.as_str());
|
||||
}
|
||||
|
||||
// update self in db
|
||||
site.crawled = true;
|
||||
site.status_code = code.as_u16();
|
||||
Website::store_all(vec![site.clone()], &db).await;
|
||||
} else {
|
||||
error!("File failed to cooperate: {:?}", path);
|
||||
}
|
||||
} else {
|
||||
error!("Failed to get: {}", &site.site);
|
||||
}
|
||||
|
||||
// process_span.end();
|
||||
BEING_PROCESSED.add(-1, &[]);
|
||||
}
|
||||
|
||||
fn load_tracing(config: &Config) {
|
||||
// Send spans to Alloy (which will send them to Tempo)
|
||||
let otlp_span = opentelemetry_otlp::SpanExporter::builder()
|
||||
.with_tonic()
|
||||
.with_endpoint(config.tracing_endpoint.clone())
|
||||
.build()
|
||||
.unwrap();
|
||||
let tracer_provider = opentelemetry_sdk::trace::SdkTracerProvider::builder()
|
||||
.with_simple_exporter(otlp_span)
|
||||
.build();
|
||||
|
||||
global::set_tracer_provider(tracer_provider);
|
||||
}
|
||||
|
||||
fn load_logging(config: &Config) {
|
||||
// let otlp_log = opentelemetry_otlp::LogExporter::builder()
|
||||
// .with_tonic()
|
||||
// .with_endpoint(endpoint)
|
||||
// .build()
|
||||
// .unwrap();
|
||||
// let logger_provider = opentelemetry_sdk::logs::SdkLoggerProvider::builder()
|
||||
// .with_simple_exporter(otlp_log)
|
||||
// .build();
|
||||
let writer = std::fs::OpenOptions::new()
|
||||
.append(true)
|
||||
.create(true)
|
||||
.open(config.log_file.clone())
|
||||
.open("./docker/logs/tracing.log")
|
||||
.expect("Couldn't make log file!");
|
||||
|
||||
let filter = EnvFilter::builder()
|
||||
@@ -325,21 +70,220 @@ fn load_logging(config: &Config) {
|
||||
);
|
||||
|
||||
tracing::subscriber::set_global_default(registry).expect("Failed to set default subscriber");
|
||||
}
|
||||
|
||||
fn load_metrics(config: &Config) {
|
||||
// Send metrics to Prometheus
|
||||
let otlp_metrics = opentelemetry_otlp::MetricExporter::builder()
|
||||
.with_http()
|
||||
.with_protocol(Protocol::HttpBinary)
|
||||
.with_endpoint(config.metrics_endpoint.clone())
|
||||
let builder = PrometheusBuilder::new();
|
||||
builder
|
||||
.with_http_listener(std::net::SocketAddr::new(
|
||||
IpAddr::V4(Ipv4Addr::from_octets([0, 0, 0, 0])),
|
||||
2500,
|
||||
))
|
||||
.install()
|
||||
.expect("failed to install recorder/exporter");
|
||||
|
||||
info!("Starting...");
|
||||
|
||||
// When getting uncrawled pages, name must contain this variable. "" will effectively get ignored.
|
||||
// let crawl_filter = "en.wikipedia.org/";
|
||||
// let budget = 50;
|
||||
let mut crawled = 0;
|
||||
|
||||
let mut file = File::open("./Crawler.toml").expect("Failed to read Crawler.toml");
|
||||
let mut buf = String::new();
|
||||
let _ = file.read_to_string(&mut buf);
|
||||
|
||||
let config: Config = toml::from_str(&buf).expect("Failed to parse Crawler.toml");
|
||||
let starting_url = &config.start_url;
|
||||
|
||||
let db = connect(&config)
|
||||
.await
|
||||
.expect("Failed to connect to surreal, aborting.");
|
||||
|
||||
let reqwest = reqwest::Client::builder()
|
||||
// .use_rustls_tls()
|
||||
.gzip(true)
|
||||
.build()
|
||||
.unwrap();
|
||||
let metrics_provider = opentelemetry_sdk::metrics::SdkMeterProvider::builder()
|
||||
.with_periodic_exporter(otlp_metrics) // default delay is 60s, turn down to like 15
|
||||
.build();
|
||||
.expect("Failed to build reqwest client.");
|
||||
|
||||
global::set_meter_provider(metrics_provider);
|
||||
// Kick off the whole machine - This Website object doesn't matter, it's just to allow for
|
||||
// get() to work.
|
||||
let span = trace_span!("Pre-Loop");
|
||||
let pre_loop_span = span.enter();
|
||||
// Download the site
|
||||
let site = Website::new(starting_url, false);
|
||||
process(site, db.clone(), reqwest.clone()).await;
|
||||
|
||||
drop(pre_loop_span);
|
||||
|
||||
let span = trace_span!("Loop");
|
||||
let span = span.enter();
|
||||
while crawled < config.budget {
|
||||
let uncrawled =
|
||||
get_uncrawled_links(&db, config.budget - crawled, config.crawl_filter.clone(), &config).await;
|
||||
if uncrawled.is_empty() {
|
||||
info!("Had more budget but finished crawling everything.");
|
||||
return;
|
||||
}
|
||||
|
||||
{
|
||||
let mut futures = JoinSet::new();
|
||||
for site in uncrawled {
|
||||
gauge!(BEING_PROCESSED).increment(1);
|
||||
futures.spawn(process(site, db.clone(), reqwest.clone()));
|
||||
// let percent = format!("{:.2}%", (crawled as f32 / budget as f32) * 100f32);
|
||||
// info!("Crawled {crawled} out of {budget} pages. ({percent})");
|
||||
}
|
||||
|
||||
let c = counter!(SITES_CRAWLED);
|
||||
// As futures complete runs code in while block
|
||||
while futures.join_next().await.is_some() {
|
||||
c.increment(1);
|
||||
gauge!(BEING_PROCESSED).decrement(1);
|
||||
crawled += 1;
|
||||
}
|
||||
}
|
||||
}
|
||||
drop(span);
|
||||
|
||||
if let Ok(mut ok) = db
|
||||
.query("count(select id from website where crawled = true)")
|
||||
.await
|
||||
{
|
||||
let res = ok.take::<Option<usize>>(0);
|
||||
if let Ok(Some(n)) = res {
|
||||
info!("Total crawled pages now equals {n}");
|
||||
}
|
||||
}
|
||||
|
||||
info!("Done");
|
||||
}
|
||||
|
||||
#[instrument(skip(db, reqwest))]
|
||||
/// Downloads and crawls and stores a webpage.
|
||||
/// It is acceptable to clone `db`, `reqwest`, and `s3` because they all use `Arc`s internally. - Noted by Oliver
|
||||
async fn process(mut site: Website, db: Surreal<Client>, reqwest: reqwest::Client) {
|
||||
// METRICS
|
||||
trace!("Process: {}", &site.site);
|
||||
// Build the request
|
||||
let request_builder = reqwest.get(site.site.to_string());
|
||||
|
||||
// METRICS
|
||||
let g = gauge!(GET_IN_FLIGHT);
|
||||
g.increment(1);
|
||||
|
||||
// Send the http request (get)
|
||||
if let Ok(response) = request_builder.send().await {
|
||||
let headers = response.headers();
|
||||
let code = response.status();
|
||||
|
||||
#[allow(non_snake_case)]
|
||||
let CT = headers.get("Content-Type");
|
||||
let ct = headers.get("content-type");
|
||||
|
||||
let ct = match (CT,ct) {
|
||||
(None, None) => {
|
||||
warn!("Server did not respond with Content-Type header. Url: {} Headers: ({:?})", site.site.to_string(), headers);
|
||||
return
|
||||
},
|
||||
(None, Some(a)) => a,
|
||||
(Some(a), None) => a,
|
||||
(Some(a), Some(_)) => a,
|
||||
};
|
||||
|
||||
// create filepath (handles / -> /index.html)
|
||||
let path = filesystem::as_path(&site.site, ct);
|
||||
|
||||
// make sure that the file is good to go
|
||||
if let Some(file) = filesystem::init(&path).await {
|
||||
// Get body from response
|
||||
// stream the response onto the disk
|
||||
let mut stream = response.bytes_stream();
|
||||
|
||||
let should_parse = path.to_string_lossy().ends_with(".html");
|
||||
let mut writer = BufWriter::new(file);
|
||||
let mut buf: Vec<u8> = Vec::new();
|
||||
|
||||
// Write file to disk
|
||||
info!("Writing at: {:?}", path);
|
||||
while let Some(data) = stream.next().await {
|
||||
match data {
|
||||
Ok(data) => {
|
||||
let _ = writer.write_all(&data).await;
|
||||
// If we are going to parse this file later, we will save it
|
||||
// into memory as well as the disk.
|
||||
// We do this because the data here might be incomplete
|
||||
if should_parse {
|
||||
data.iter().for_each(|f| buf.push(*f));
|
||||
}
|
||||
},
|
||||
Err(err) => {
|
||||
eprintln!("{}", err)
|
||||
},
|
||||
}
|
||||
}
|
||||
let _ = writer.flush();
|
||||
|
||||
|
||||
// (If needed) Parse the file
|
||||
if should_parse {
|
||||
let span = debug_span!("Should Parse");
|
||||
let enter = span.enter();
|
||||
|
||||
// Parse document and get relationships
|
||||
let sites = parser::parse(&site, &buf).await;
|
||||
// De-duplicate this list
|
||||
let prev_len = sites.len();
|
||||
let set = sites.into_iter().fold(HashSet::new(), |mut set, item| {
|
||||
set.insert(item);
|
||||
set
|
||||
});
|
||||
let de_dupe_sites: Vec<Website> = set.into_iter().collect();
|
||||
let diff = prev_len - de_dupe_sites.len();
|
||||
trace!("Saved {diff} from being entered into the db by de-duping");
|
||||
// Store all the other sites so that we can link to them.
|
||||
let _ = Website::store_all(de_dupe_sites, &db).await;
|
||||
|
||||
drop(enter);
|
||||
}
|
||||
|
||||
// METRICS
|
||||
g.decrement(1);
|
||||
counter!(GET_METRIC).increment(1);
|
||||
|
||||
// update self in db
|
||||
site.crawled = true;
|
||||
site.status_code = code.as_u16();
|
||||
Website::store_all(vec![site.clone()], &db).await;
|
||||
} else {
|
||||
error!("File failed to cooperate: {:?}", path);
|
||||
}
|
||||
|
||||
trace!("Done processing: {}", &site.site);
|
||||
} else {
|
||||
error!("Failed to get: {}", &site.site);
|
||||
}
|
||||
}
|
||||
|
||||
/// Returns uncrawled links
|
||||
#[instrument(skip(db, config))]
|
||||
async fn get_uncrawled_links(
|
||||
db: &Surreal<Client>,
|
||||
mut count: usize,
|
||||
filter: String,
|
||||
config: &Config,
|
||||
) -> Vec<Website> {
|
||||
if count > config.batch_size {
|
||||
count = config.batch_size;
|
||||
}
|
||||
|
||||
debug!("Getting {} uncrawled links", count);
|
||||
|
||||
let mut response = db
|
||||
.query("SELECT * FROM website WHERE crawled = false AND site ~ type::string($format) LIMIT $count;")
|
||||
.bind(("format", filter))
|
||||
.bind(("count", count))
|
||||
.await
|
||||
.expect("Hard-coded query failed..?");
|
||||
response
|
||||
.take(0)
|
||||
.expect("Returned websites couldn't be parsed")
|
||||
}
|
||||
|
||||
@@ -4,7 +4,7 @@ use html5ever::tokenizer::{BufferQueue, TokenizerResult};
|
||||
use html5ever::tokenizer::{StartTag, TagToken};
|
||||
use html5ever::tokenizer::{Token, TokenSink, TokenSinkResult, Tokenizer, TokenizerOpts};
|
||||
use html5ever::{local_name, tendril::*};
|
||||
use tracing::{error, instrument, trace, warn};
|
||||
use tracing::{debug, error, instrument, trace, warn};
|
||||
use url::Url;
|
||||
|
||||
use crate::db::Website;
|
||||
@@ -12,6 +12,7 @@ use crate::db::Website;
|
||||
impl TokenSink for Website {
|
||||
type Handle = Vec<Website>;
|
||||
|
||||
#[instrument(skip(token, _line_number))]
|
||||
fn process_token(&self, token: Token, _line_number: u64) -> TokenSinkResult<Self::Handle> {
|
||||
match token {
|
||||
TagToken(tag) => {
|
||||
@@ -32,13 +33,13 @@ impl TokenSink for Website {
|
||||
let attr_name = attr.name.local.to_string();
|
||||
if attr_name == "src" || attr_name == "href" || attr_name == "data"
|
||||
{
|
||||
trace!(url = self.site.as_str(),"Found `{}` in html `{}` tag", &attr.value, tag.name);
|
||||
trace!("Found `{}` in html `{}` tag", &attr.value, tag.name);
|
||||
let url = try_get_url(&self.site, &attr.value);
|
||||
|
||||
if let Some(mut parsed) = url {
|
||||
parsed.set_query(None);
|
||||
parsed.set_fragment(None);
|
||||
trace!(url = self.site.as_str(), "Final cleaned URL: `{}`", parsed.to_string());
|
||||
trace!("Final cleaned URL: `{}`", parsed.to_string());
|
||||
let web = Website::new(&parsed.to_string(), false);
|
||||
links.push(web);
|
||||
}
|
||||
@@ -59,10 +60,10 @@ impl TokenSink for Website {
|
||||
}
|
||||
}
|
||||
|
||||
#[instrument(skip(data))]
|
||||
#[instrument(skip_all)]
|
||||
/// Parses the passed site and returns all the sites it links to.
|
||||
pub async fn parse(site: &Website, data: &[u8]) -> Vec<Website> {
|
||||
trace!(url = site.site.as_str(), "Parsing {}", site.site.to_string());
|
||||
debug!("Parsing {}", site.site.to_string());
|
||||
// prep work
|
||||
let mut other_sites: Vec<Website> = Vec::new();
|
||||
|
||||
@@ -87,7 +88,7 @@ pub async fn parse(site: &Website, data: &[u8]) -> Vec<Website> {
|
||||
assert!(token_buffer.is_empty());
|
||||
tokenizer.end();
|
||||
} else {
|
||||
warn!(url = site.site.as_str(), "Tendril failed to parse on: {}", site.site.to_string());
|
||||
warn!("Tendril failed to parse on: {}", site.site.to_string());
|
||||
}
|
||||
|
||||
other_sites
|
||||
@@ -99,7 +100,7 @@ fn try_get_url(parent: &Url, link: &str) -> Option<Url> {
|
||||
Ok(ok) => Some(ok),
|
||||
Err(e) => {
|
||||
if link.starts_with('#') {
|
||||
trace!(url = parent.as_str(), "Rejecting # url");
|
||||
trace!("Rejecting # url");
|
||||
None
|
||||
} else if link.starts_with("//") {
|
||||
// if a url starts with "//" is assumed that it will adopt
|
||||
@@ -107,7 +108,7 @@ fn try_get_url(parent: &Url, link: &str) -> Option<Url> {
|
||||
// https://stackoverflow.com/questions/9646407/two-forward-slashes-in-a-url-src-href-attribute
|
||||
let scheme = parent.scheme();
|
||||
|
||||
match Url::parse(&format!("{scheme}://{link}")) {
|
||||
match Url::parse(&format!("{scheme}://{}", link)) {
|
||||
Ok(url) => Some(url),
|
||||
Err(err) => {
|
||||
error!("Failed parsing relative scheme url: {}", err);
|
||||
@@ -115,7 +116,7 @@ fn try_get_url(parent: &Url, link: &str) -> Option<Url> {
|
||||
}
|
||||
}
|
||||
} else {
|
||||
// # This is some sort of relative url, gonna try patching it up into an absolute
|
||||
// # This is some sort of realative url, gonna try patching it up into an absolute
|
||||
// url
|
||||
match e {
|
||||
url::ParseError::RelativeUrlWithoutBase => {
|
||||
@@ -127,7 +128,7 @@ fn try_get_url(parent: &Url, link: &str) -> Option<Url> {
|
||||
let url = origin.clone() + link;
|
||||
|
||||
if let Ok(url) = Url::parse(&url) {
|
||||
trace!(url = parent.as_str(), "Built `{url}` from `{origin} + `{}`", link.to_string());
|
||||
trace!("Built `{url}` from `{origin} + `{}`", link.to_string());
|
||||
Some(url)
|
||||
} else {
|
||||
error!(
|
||||
|
||||
@@ -4,13 +4,6 @@ DEFINE FIELD IF NOT EXISTS site ON TABLE website TYPE string;
|
||||
DEFINE INDEX IF NOT EXISTS idx ON TABLE website COLUMNS site UNIQUE;
|
||||
|
||||
DEFINE FIELD IF NOT EXISTS crawled ON TABLE website TYPE bool;
|
||||
DEFINE FIELD IF NOT EXISTS processing ON TABLE website TYPE bool DEFAULT false;
|
||||
|
||||
DEFINE FIELD IF NOT EXISTS accessed_at ON TABLE website VALUE time::now();
|
||||
DEFINE FIELD IF NOT EXISTS first_accessed_at ON TABLE website VALUE time::now();
|
||||
|
||||
DEFINE FUNCTION OVERWRITE fn::get_next($filter: string) {
|
||||
LET $site = SELECT * FROM ONLY website WHERE crawled = false AND processing = false AND site ~ type::string($filter) LIMIT 1;
|
||||
UPDATE $site.id SET processing = true;
|
||||
RETURN $site
|
||||
};
|
||||
|
||||
Reference in New Issue
Block a user