Compare commits
6 Commits
135a7e4957
...
71b7b2d7bc
Author | SHA1 | Date | |
---|---|---|---|
71b7b2d7bc | |||
bac3cd9d1d | |||
1f6a0acce3 | |||
53dbf53ab9 | |||
0477bb26e4 | |||
6409baaffb |
8
.vscode/settings.json
vendored
Normal file
8
.vscode/settings.json
vendored
Normal file
@ -0,0 +1,8 @@
|
||||
{
|
||||
"cSpell.words": [
|
||||
"creds",
|
||||
"reqwest",
|
||||
"rustls",
|
||||
"surql"
|
||||
]
|
||||
}
|
10
Crawler.toml
10
Crawler.toml
@ -3,14 +3,14 @@ surreal_url = "localhost:8000"
|
||||
surreal_username = "root"
|
||||
surreal_password = "root"
|
||||
surreal_ns = "test"
|
||||
surreal_db = "v1.12"
|
||||
surreal_db = "v1.15.4"
|
||||
|
||||
# Minio config
|
||||
s3_bucket = "v1.12"
|
||||
s3_bucket = "v1.15.4"
|
||||
s3_url = "http://localhost:9000"
|
||||
s3_access_key = "jLDPKGuu513VENc8kJwX"
|
||||
s3_secret_key = "4T1nymEzsGYOlKSAb1WX7V3stnQn9a5ZoTQjDfcL"
|
||||
s3_access_key = "3ptjsHhRHCHlpCmgFy9n"
|
||||
s3_secret_key = "68CmV07YExeCxb8kJhosSauEizj5CAE7PINZIfQz"
|
||||
|
||||
# Crawler config
|
||||
crawl_filter = "en.wikipedia.com"
|
||||
budget = 200
|
||||
budget = 1000
|
||||
|
@ -8,4 +8,5 @@ Crawls sites saving all the found links to a surrealdb database. It then proceed
|
||||
- [ ] Domain filtering - prevent the crawler from going on alternate versions of wikipedia.
|
||||
- [ ] Conditionally save content - based on filename or file contents
|
||||
- [ ] GUI / TUI ?
|
||||
- [ ] Better asynchronous getting of the sites. Currently it all happens serially.
|
||||
- [ ] Better asynchronous getting of the sites. Currently it all happens serially.3/19/25: Took 20min to crawl 100 pages
|
||||
This ment we stored 100 pages, 142,997 urls, and 1,425,798 links between the two.
|
||||
|
@ -22,6 +22,7 @@ services:
|
||||
environment:
|
||||
- MINIO_ROOT_USER=root
|
||||
- MINIO_ROOT_PASSWORD=an8charpassword
|
||||
- MINIO_PROMETHEUS_AUTH_TYPE=public
|
||||
volumes:
|
||||
- minio_storage:/data
|
||||
command:
|
||||
@ -66,6 +67,7 @@ services:
|
||||
image: grafana/grafana:latest
|
||||
volumes:
|
||||
- ./grafana.yaml:/etc/grafana/provisioning/datasources/datasources.yaml
|
||||
- ./dashboards:/var/lib/grafana/dashboards
|
||||
- grafana_storage:/var/lib/grafana
|
||||
environment:
|
||||
- GF_AUTH_ANONYMOUS_ENABLED=true
|
||||
@ -80,4 +82,4 @@ volumes:
|
||||
grafana_storage:
|
||||
alloy_storage:
|
||||
surrealdb_storage:
|
||||
minio_storage:
|
||||
minio_storage:
|
||||
|
648
docker/dashboards/crawler-dashboard.json
Normal file
648
docker/dashboards/crawler-dashboard.json
Normal file
@ -0,0 +1,648 @@
|
||||
{
|
||||
"__inputs": [
|
||||
{
|
||||
"name": "DS_PROMETHEUS",
|
||||
"label": "Prometheus",
|
||||
"description": "",
|
||||
"type": "datasource",
|
||||
"pluginId": "prometheus",
|
||||
"pluginName": "Prometheus"
|
||||
},
|
||||
{
|
||||
"name": "DS_LOKI",
|
||||
"label": "Loki",
|
||||
"description": "",
|
||||
"type": "datasource",
|
||||
"pluginId": "loki",
|
||||
"pluginName": "Loki"
|
||||
}
|
||||
],
|
||||
"__elements": {},
|
||||
"__requires": [
|
||||
{
|
||||
"type": "grafana",
|
||||
"id": "grafana",
|
||||
"name": "Grafana",
|
||||
"version": "11.3.1"
|
||||
},
|
||||
{
|
||||
"type": "panel",
|
||||
"id": "logs",
|
||||
"name": "Logs",
|
||||
"version": ""
|
||||
},
|
||||
{
|
||||
"type": "datasource",
|
||||
"id": "loki",
|
||||
"name": "Loki",
|
||||
"version": "1.0.0"
|
||||
},
|
||||
{
|
||||
"type": "datasource",
|
||||
"id": "prometheus",
|
||||
"name": "Prometheus",
|
||||
"version": "1.0.0"
|
||||
},
|
||||
{
|
||||
"type": "panel",
|
||||
"id": "stat",
|
||||
"name": "Stat",
|
||||
"version": ""
|
||||
},
|
||||
{
|
||||
"type": "panel",
|
||||
"id": "timeseries",
|
||||
"name": "Time series",
|
||||
"version": ""
|
||||
}
|
||||
],
|
||||
"annotations": {
|
||||
"list": [
|
||||
{
|
||||
"builtIn": 1,
|
||||
"datasource": {
|
||||
"type": "grafana",
|
||||
"uid": "-- Grafana --"
|
||||
},
|
||||
"enable": true,
|
||||
"hide": true,
|
||||
"iconColor": "rgba(0, 211, 255, 1)",
|
||||
"name": "Annotations & Alerts",
|
||||
"type": "dashboard"
|
||||
}
|
||||
]
|
||||
},
|
||||
"editable": true,
|
||||
"fiscalYearStartMonth": 0,
|
||||
"graphTooltip": 0,
|
||||
"id": null,
|
||||
"links": [],
|
||||
"panels": [
|
||||
{
|
||||
"datasource": {
|
||||
"type": "prometheus",
|
||||
"uid": "${DS_PROMETHEUS}"
|
||||
},
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": {
|
||||
"mode": "palette-classic"
|
||||
},
|
||||
"custom": {
|
||||
"axisBorderShow": false,
|
||||
"axisCenteredZero": false,
|
||||
"axisColorMode": "text",
|
||||
"axisLabel": "",
|
||||
"axisPlacement": "auto",
|
||||
"barAlignment": 0,
|
||||
"barWidthFactor": 0.6,
|
||||
"drawStyle": "line",
|
||||
"fillOpacity": 0,
|
||||
"gradientMode": "none",
|
||||
"hideFrom": {
|
||||
"legend": false,
|
||||
"tooltip": false,
|
||||
"viz": false
|
||||
},
|
||||
"insertNulls": false,
|
||||
"lineInterpolation": "linear",
|
||||
"lineWidth": 1,
|
||||
"pointSize": 5,
|
||||
"scaleDistribution": {
|
||||
"type": "linear"
|
||||
},
|
||||
"showPoints": "auto",
|
||||
"spanNulls": 300000,
|
||||
"stacking": {
|
||||
"group": "A",
|
||||
"mode": "none"
|
||||
},
|
||||
"thresholdsStyle": {
|
||||
"mode": "off"
|
||||
}
|
||||
},
|
||||
"mappings": [],
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{
|
||||
"color": "green",
|
||||
"value": null
|
||||
},
|
||||
{
|
||||
"color": "red",
|
||||
"value": 80
|
||||
}
|
||||
]
|
||||
}
|
||||
},
|
||||
"overrides": []
|
||||
},
|
||||
"gridPos": {
|
||||
"h": 8,
|
||||
"w": 8,
|
||||
"x": 0,
|
||||
"y": 0
|
||||
},
|
||||
"id": 5,
|
||||
"options": {
|
||||
"legend": {
|
||||
"calcs": [],
|
||||
"displayMode": "list",
|
||||
"placement": "bottom",
|
||||
"showLegend": true
|
||||
},
|
||||
"tooltip": {
|
||||
"mode": "single",
|
||||
"sort": "none"
|
||||
}
|
||||
},
|
||||
"pluginVersion": "11.3.1",
|
||||
"targets": [
|
||||
{
|
||||
"datasource": {
|
||||
"type": "prometheus",
|
||||
"uid": "${DS_PROMETHEUS}"
|
||||
},
|
||||
"disableTextWrap": false,
|
||||
"editorMode": "builder",
|
||||
"expr": "surql_trips",
|
||||
"fullMetaSearch": false,
|
||||
"includeNullMetadata": true,
|
||||
"legendFormat": "Trips to Surreal",
|
||||
"range": true,
|
||||
"refId": "A",
|
||||
"useBackend": false
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "prometheus",
|
||||
"uid": "${DS_PROMETHEUS}"
|
||||
},
|
||||
"disableTextWrap": false,
|
||||
"editorMode": "builder",
|
||||
"expr": "s3_trips",
|
||||
"fullMetaSearch": false,
|
||||
"hide": false,
|
||||
"includeNullMetadata": true,
|
||||
"instant": false,
|
||||
"legendFormat": "Trips to S3",
|
||||
"range": true,
|
||||
"refId": "B",
|
||||
"useBackend": false
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "prometheus",
|
||||
"uid": "${DS_PROMETHEUS}"
|
||||
},
|
||||
"disableTextWrap": false,
|
||||
"editorMode": "builder",
|
||||
"expr": "pages_crawled",
|
||||
"fullMetaSearch": false,
|
||||
"hide": false,
|
||||
"includeNullMetadata": true,
|
||||
"instant": false,
|
||||
"legendFormat": "total crawled",
|
||||
"range": true,
|
||||
"refId": "C",
|
||||
"useBackend": false
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "prometheus",
|
||||
"uid": "${DS_PROMETHEUS}"
|
||||
},
|
||||
"disableTextWrap": false,
|
||||
"editorMode": "builder",
|
||||
"expr": "pages_being_processed",
|
||||
"fullMetaSearch": false,
|
||||
"hide": false,
|
||||
"includeNullMetadata": true,
|
||||
"instant": false,
|
||||
"legendFormat": "Pages being processed",
|
||||
"range": true,
|
||||
"refId": "E",
|
||||
"useBackend": false
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "prometheus",
|
||||
"uid": "${DS_PROMETHEUS}"
|
||||
},
|
||||
"disableTextWrap": false,
|
||||
"editorMode": "builder",
|
||||
"expr": "gets_in_flight",
|
||||
"fullMetaSearch": false,
|
||||
"hide": false,
|
||||
"includeNullMetadata": true,
|
||||
"instant": false,
|
||||
"legendFormat": "__auto",
|
||||
"range": true,
|
||||
"refId": "D",
|
||||
"useBackend": false
|
||||
}
|
||||
],
|
||||
"title": "Crawler stats",
|
||||
"type": "timeseries"
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "prometheus",
|
||||
"uid": "${DS_PROMETHEUS}"
|
||||
},
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": {
|
||||
"mode": "palette-classic"
|
||||
},
|
||||
"custom": {
|
||||
"axisBorderShow": false,
|
||||
"axisCenteredZero": false,
|
||||
"axisColorMode": "text",
|
||||
"axisLabel": "",
|
||||
"axisPlacement": "auto",
|
||||
"barAlignment": 0,
|
||||
"barWidthFactor": 0.6,
|
||||
"drawStyle": "line",
|
||||
"fillOpacity": 0,
|
||||
"gradientMode": "none",
|
||||
"hideFrom": {
|
||||
"legend": false,
|
||||
"tooltip": false,
|
||||
"viz": false
|
||||
},
|
||||
"insertNulls": false,
|
||||
"lineInterpolation": "linear",
|
||||
"lineWidth": 1,
|
||||
"pointSize": 5,
|
||||
"scaleDistribution": {
|
||||
"type": "linear"
|
||||
},
|
||||
"showPoints": "auto",
|
||||
"spanNulls": 300000,
|
||||
"stacking": {
|
||||
"group": "A",
|
||||
"mode": "none"
|
||||
},
|
||||
"thresholdsStyle": {
|
||||
"mode": "off"
|
||||
}
|
||||
},
|
||||
"mappings": [],
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{
|
||||
"color": "green",
|
||||
"value": null
|
||||
},
|
||||
{
|
||||
"color": "red",
|
||||
"value": 80
|
||||
}
|
||||
]
|
||||
}
|
||||
},
|
||||
"overrides": []
|
||||
},
|
||||
"gridPos": {
|
||||
"h": 8,
|
||||
"w": 9,
|
||||
"x": 8,
|
||||
"y": 0
|
||||
},
|
||||
"id": 6,
|
||||
"options": {
|
||||
"legend": {
|
||||
"calcs": [],
|
||||
"displayMode": "list",
|
||||
"placement": "bottom",
|
||||
"showLegend": true
|
||||
},
|
||||
"tooltip": {
|
||||
"mode": "single",
|
||||
"sort": "none"
|
||||
}
|
||||
},
|
||||
"pluginVersion": "11.3.1",
|
||||
"targets": [
|
||||
{
|
||||
"datasource": {
|
||||
"type": "prometheus",
|
||||
"uid": "${DS_PROMETHEUS}"
|
||||
},
|
||||
"disableTextWrap": false,
|
||||
"editorMode": "builder",
|
||||
"expr": "surql_trips",
|
||||
"fullMetaSearch": false,
|
||||
"includeNullMetadata": true,
|
||||
"legendFormat": "Trips to Surreal",
|
||||
"range": true,
|
||||
"refId": "A",
|
||||
"useBackend": false
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "prometheus",
|
||||
"uid": "${DS_PROMETHEUS}"
|
||||
},
|
||||
"disableTextWrap": false,
|
||||
"editorMode": "builder",
|
||||
"expr": "surql_link_calls",
|
||||
"fullMetaSearch": false,
|
||||
"hide": false,
|
||||
"includeNullMetadata": true,
|
||||
"instant": false,
|
||||
"legendFormat": "link calls",
|
||||
"range": true,
|
||||
"refId": "B",
|
||||
"useBackend": false
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "prometheus",
|
||||
"uid": "${DS_PROMETHEUS}"
|
||||
},
|
||||
"disableTextWrap": false,
|
||||
"editorMode": "builder",
|
||||
"expr": "surql_store_calls",
|
||||
"fullMetaSearch": false,
|
||||
"hide": false,
|
||||
"includeNullMetadata": true,
|
||||
"instant": false,
|
||||
"legendFormat": "store calls",
|
||||
"range": true,
|
||||
"refId": "C",
|
||||
"useBackend": false
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "prometheus",
|
||||
"uid": "${DS_PROMETHEUS}"
|
||||
},
|
||||
"disableTextWrap": false,
|
||||
"editorMode": "builder",
|
||||
"expr": "pages_being_processed",
|
||||
"fullMetaSearch": false,
|
||||
"hide": false,
|
||||
"includeNullMetadata": true,
|
||||
"instant": false,
|
||||
"legendFormat": "Pages being processed",
|
||||
"range": true,
|
||||
"refId": "E",
|
||||
"useBackend": false
|
||||
}
|
||||
],
|
||||
"title": "Surreal stats",
|
||||
"type": "timeseries"
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "prometheus",
|
||||
"uid": "${DS_PROMETHEUS}"
|
||||
},
|
||||
"description": "This is across all threads, so this isn't wall clock time",
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": {
|
||||
"mode": "thresholds"
|
||||
},
|
||||
"mappings": [],
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{
|
||||
"color": "green",
|
||||
"value": null
|
||||
}
|
||||
]
|
||||
},
|
||||
"unit": "ms"
|
||||
},
|
||||
"overrides": []
|
||||
},
|
||||
"gridPos": {
|
||||
"h": 8,
|
||||
"w": 7,
|
||||
"x": 17,
|
||||
"y": 0
|
||||
},
|
||||
"id": 7,
|
||||
"options": {
|
||||
"colorMode": "value",
|
||||
"graphMode": "area",
|
||||
"justifyMode": "auto",
|
||||
"orientation": "auto",
|
||||
"percentChangeColorMode": "standard",
|
||||
"reduceOptions": {
|
||||
"calcs": [
|
||||
"lastNotNull"
|
||||
],
|
||||
"fields": "",
|
||||
"values": false
|
||||
},
|
||||
"showPercentChange": false,
|
||||
"textMode": "auto",
|
||||
"wideLayout": true
|
||||
},
|
||||
"pluginVersion": "11.3.1",
|
||||
"targets": [
|
||||
{
|
||||
"datasource": {
|
||||
"type": "prometheus",
|
||||
"uid": "${DS_PROMETHEUS}"
|
||||
},
|
||||
"disableTextWrap": false,
|
||||
"editorMode": "builder",
|
||||
"expr": "surql_lock_waiting_ms",
|
||||
"fullMetaSearch": false,
|
||||
"includeNullMetadata": true,
|
||||
"legendFormat": "__auto",
|
||||
"range": true,
|
||||
"refId": "A",
|
||||
"useBackend": false
|
||||
}
|
||||
],
|
||||
"title": "Time spend waiting on lock",
|
||||
"type": "stat"
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"gridPos": {
|
||||
"h": 18,
|
||||
"w": 24,
|
||||
"x": 0,
|
||||
"y": 8
|
||||
},
|
||||
"id": 1,
|
||||
"options": {
|
||||
"dedupStrategy": "none",
|
||||
"enableLogDetails": true,
|
||||
"prettifyLogMessage": false,
|
||||
"showCommonLabels": false,
|
||||
"showLabels": false,
|
||||
"showTime": false,
|
||||
"sortOrder": "Descending",
|
||||
"wrapLogMessage": false
|
||||
},
|
||||
"pluginVersion": "11.3.1",
|
||||
"targets": [
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"editorMode": "code",
|
||||
"expr": "{filename=\"/tmp/alloy-logs/tracing.log\"} | json | level = `ERROR` | line_format \"{{.threadId}} {{.filename_extracted}}:{{.line_number}} {{.fields_message}}\"",
|
||||
"queryType": "range",
|
||||
"refId": "A"
|
||||
}
|
||||
],
|
||||
"title": "Errors",
|
||||
"type": "logs"
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"gridPos": {
|
||||
"h": 8,
|
||||
"w": 12,
|
||||
"x": 0,
|
||||
"y": 26
|
||||
},
|
||||
"id": 2,
|
||||
"options": {
|
||||
"dedupStrategy": "none",
|
||||
"enableLogDetails": true,
|
||||
"prettifyLogMessage": false,
|
||||
"showCommonLabels": false,
|
||||
"showLabels": false,
|
||||
"showTime": false,
|
||||
"sortOrder": "Descending",
|
||||
"wrapLogMessage": false
|
||||
},
|
||||
"pluginVersion": "11.3.1",
|
||||
"targets": [
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"editorMode": "code",
|
||||
"expr": "{filename=\"/tmp/alloy-logs/tracing.log\"} | json | level = `DEBUG` | line_format \"{{.fields_message}}\"",
|
||||
"queryType": "range",
|
||||
"refId": "A"
|
||||
}
|
||||
],
|
||||
"title": "Debug",
|
||||
"type": "logs"
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"gridPos": {
|
||||
"h": 16,
|
||||
"w": 12,
|
||||
"x": 12,
|
||||
"y": 26
|
||||
},
|
||||
"id": 4,
|
||||
"options": {
|
||||
"dedupStrategy": "none",
|
||||
"enableLogDetails": true,
|
||||
"prettifyLogMessage": false,
|
||||
"showCommonLabels": false,
|
||||
"showLabels": false,
|
||||
"showTime": false,
|
||||
"sortOrder": "Descending",
|
||||
"wrapLogMessage": false
|
||||
},
|
||||
"pluginVersion": "11.3.1",
|
||||
"targets": [
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"editorMode": "code",
|
||||
"expr": "{filename=\"/tmp/alloy-logs/tracing.log\"} | json | level = `TRACE` | line_format \"{{.fields_message}}\"",
|
||||
"queryType": "range",
|
||||
"refId": "A"
|
||||
}
|
||||
],
|
||||
"title": "Trace",
|
||||
"type": "logs"
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"gridPos": {
|
||||
"h": 8,
|
||||
"w": 12,
|
||||
"x": 0,
|
||||
"y": 34
|
||||
},
|
||||
"id": 3,
|
||||
"options": {
|
||||
"dedupStrategy": "none",
|
||||
"enableLogDetails": true,
|
||||
"prettifyLogMessage": false,
|
||||
"showCommonLabels": false,
|
||||
"showLabels": false,
|
||||
"showTime": false,
|
||||
"sortOrder": "Descending",
|
||||
"wrapLogMessage": false
|
||||
},
|
||||
"pluginVersion": "11.3.1",
|
||||
"targets": [
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"editorMode": "code",
|
||||
"expr": "{filename=\"/tmp/alloy-logs/tracing.log\"} | json | level = `WARN` | line_format \"{{.fields_message}}\"",
|
||||
"queryType": "range",
|
||||
"refId": "A"
|
||||
}
|
||||
],
|
||||
"title": "Warnings",
|
||||
"type": "logs"
|
||||
}
|
||||
],
|
||||
"schemaVersion": 40,
|
||||
"tags": [],
|
||||
"templating": {
|
||||
"list": [
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "P8E80F9AEF21F6940"
|
||||
},
|
||||
"filters": [],
|
||||
"name": "Filters",
|
||||
"type": "adhoc"
|
||||
}
|
||||
]
|
||||
},
|
||||
"time": {
|
||||
"from": "now-5m",
|
||||
"to": "now"
|
||||
},
|
||||
"timepicker": {},
|
||||
"timezone": "browser",
|
||||
"title": "Crawler",
|
||||
"uid": "ceg90x34pqgowd",
|
||||
"version": 21,
|
||||
"weekStart": ""
|
||||
}
|
@ -1,223 +0,0 @@
|
||||
{
|
||||
"__inputs": [
|
||||
{
|
||||
"name": "DS_LOKI",
|
||||
"label": "Loki",
|
||||
"description": "",
|
||||
"type": "datasource",
|
||||
"pluginId": "loki",
|
||||
"pluginName": "Loki"
|
||||
}
|
||||
],
|
||||
"__elements": {},
|
||||
"__requires": [
|
||||
{
|
||||
"type": "grafana",
|
||||
"id": "grafana",
|
||||
"name": "Grafana",
|
||||
"version": "11.3.1"
|
||||
},
|
||||
{
|
||||
"type": "panel",
|
||||
"id": "logs",
|
||||
"name": "Logs",
|
||||
"version": ""
|
||||
},
|
||||
{
|
||||
"type": "datasource",
|
||||
"id": "loki",
|
||||
"name": "Loki",
|
||||
"version": "1.0.0"
|
||||
}
|
||||
],
|
||||
"annotations": {
|
||||
"list": [
|
||||
{
|
||||
"builtIn": 1,
|
||||
"datasource": {
|
||||
"type": "grafana",
|
||||
"uid": "-- Grafana --"
|
||||
},
|
||||
"enable": true,
|
||||
"hide": true,
|
||||
"iconColor": "rgba(0, 211, 255, 1)",
|
||||
"name": "Annotations & Alerts",
|
||||
"type": "dashboard"
|
||||
}
|
||||
]
|
||||
},
|
||||
"editable": true,
|
||||
"fiscalYearStartMonth": 0,
|
||||
"graphTooltip": 0,
|
||||
"id": null,
|
||||
"links": [],
|
||||
"panels": [
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"gridPos": {
|
||||
"h": 8,
|
||||
"w": 12,
|
||||
"x": 0,
|
||||
"y": 0
|
||||
},
|
||||
"id": 1,
|
||||
"options": {
|
||||
"dedupStrategy": "none",
|
||||
"enableLogDetails": true,
|
||||
"prettifyLogMessage": false,
|
||||
"showCommonLabels": false,
|
||||
"showLabels": false,
|
||||
"showTime": false,
|
||||
"sortOrder": "Descending",
|
||||
"wrapLogMessage": false
|
||||
},
|
||||
"pluginVersion": "11.3.1",
|
||||
"targets": [
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"editorMode": "code",
|
||||
"expr": "{filename=\"/tmp/alloy-logs/tracing.log\"} | json | level = `ERROR` | line_format \"{{.fields_message}}\"",
|
||||
"queryType": "range",
|
||||
"refId": "A"
|
||||
}
|
||||
],
|
||||
"title": "Errors",
|
||||
"type": "logs"
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"gridPos": {
|
||||
"h": 8,
|
||||
"w": 12,
|
||||
"x": 12,
|
||||
"y": 0
|
||||
},
|
||||
"id": 3,
|
||||
"options": {
|
||||
"dedupStrategy": "none",
|
||||
"enableLogDetails": true,
|
||||
"prettifyLogMessage": false,
|
||||
"showCommonLabels": false,
|
||||
"showLabels": false,
|
||||
"showTime": false,
|
||||
"sortOrder": "Descending",
|
||||
"wrapLogMessage": false
|
||||
},
|
||||
"pluginVersion": "11.3.1",
|
||||
"targets": [
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"editorMode": "code",
|
||||
"expr": "{filename=\"/tmp/alloy-logs/tracing.log\"} | json | level = `WARN` | line_format \"{{.fields_message}}\"",
|
||||
"queryType": "range",
|
||||
"refId": "A"
|
||||
}
|
||||
],
|
||||
"title": "Warnings",
|
||||
"type": "logs"
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"gridPos": {
|
||||
"h": 8,
|
||||
"w": 12,
|
||||
"x": 0,
|
||||
"y": 8
|
||||
},
|
||||
"id": 2,
|
||||
"options": {
|
||||
"dedupStrategy": "none",
|
||||
"enableLogDetails": true,
|
||||
"prettifyLogMessage": false,
|
||||
"showCommonLabels": false,
|
||||
"showLabels": false,
|
||||
"showTime": false,
|
||||
"sortOrder": "Descending",
|
||||
"wrapLogMessage": false
|
||||
},
|
||||
"pluginVersion": "11.3.1",
|
||||
"targets": [
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"editorMode": "code",
|
||||
"expr": "{filename=\"/tmp/alloy-logs/tracing.log\"} | json | level = `DEBUG` | line_format \"{{.fields_message}}\"",
|
||||
"queryType": "range",
|
||||
"refId": "A"
|
||||
}
|
||||
],
|
||||
"title": "Debug",
|
||||
"type": "logs"
|
||||
},
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"gridPos": {
|
||||
"h": 8,
|
||||
"w": 12,
|
||||
"x": 12,
|
||||
"y": 8
|
||||
},
|
||||
"id": 4,
|
||||
"options": {
|
||||
"dedupStrategy": "none",
|
||||
"enableLogDetails": true,
|
||||
"prettifyLogMessage": false,
|
||||
"showCommonLabels": false,
|
||||
"showLabels": false,
|
||||
"showTime": false,
|
||||
"sortOrder": "Descending",
|
||||
"wrapLogMessage": false
|
||||
},
|
||||
"pluginVersion": "11.3.1",
|
||||
"targets": [
|
||||
{
|
||||
"datasource": {
|
||||
"type": "loki",
|
||||
"uid": "${DS_LOKI}"
|
||||
},
|
||||
"editorMode": "code",
|
||||
"expr": "{filename=\"/tmp/alloy-logs/tracing.log\"} | json | level = `TRACE` | line_format \"{{.fields_message}}\"",
|
||||
"queryType": "range",
|
||||
"refId": "A"
|
||||
}
|
||||
],
|
||||
"title": "Trace",
|
||||
"type": "logs"
|
||||
}
|
||||
],
|
||||
"schemaVersion": 40,
|
||||
"tags": [],
|
||||
"templating": {
|
||||
"list": []
|
||||
},
|
||||
"time": {
|
||||
"from": "now-6h",
|
||||
"to": "now"
|
||||
},
|
||||
"timepicker": {},
|
||||
"timezone": "browser",
|
||||
"title": "New dashboard",
|
||||
"uid": "ceg90x34pqgowd",
|
||||
"version": 4,
|
||||
"weekStart": ""
|
||||
}
|
@ -7,10 +7,14 @@ scrape_configs:
|
||||
static_configs:
|
||||
# change this your machine's ip, localhost won't work
|
||||
# because localhost refers to the docker container.
|
||||
- targets: ['192.168.8.209:2500']
|
||||
- targets: ['172.20.239.48:2500']
|
||||
- job_name: loki
|
||||
static_configs:
|
||||
- targets: ['loki:3100']
|
||||
- job_name: prometheus
|
||||
static_configs:
|
||||
- targets: ['localhost:9090']
|
||||
- job_name: minio
|
||||
metrics_path: /minio/v2/metrics/cluster
|
||||
static_configs:
|
||||
- targets: ['minio:9000']
|
||||
|
211
src/db.rs
211
src/db.rs
@ -1,33 +1,58 @@
|
||||
use std::fmt::Debug;
|
||||
use metrics::counter;
|
||||
use serde::{Deserialize, Serialize};
|
||||
use surrealdb::{
|
||||
engine::remote::ws::{Client, Ws}, error::Db, opt::auth::Root, sql::Thing, Response, Surreal
|
||||
use base64::{
|
||||
alphabet,
|
||||
engine::{self, general_purpose},
|
||||
Engine,
|
||||
};
|
||||
use metrics::counter;
|
||||
use serde::{ser::SerializeStruct, Deserialize, Serialize};
|
||||
use std::{fmt::Debug, sync::LazyLock, time::Instant};
|
||||
use surrealdb::{
|
||||
engine::remote::ws::{Client, Ws},
|
||||
opt::auth::Root,
|
||||
sql::Thing,
|
||||
Error::Api,
|
||||
Response, Surreal,
|
||||
};
|
||||
use tokio::sync::Mutex;
|
||||
use tracing::{error, instrument, trace, warn};
|
||||
use url::Url;
|
||||
|
||||
use crate::{Config, Timer};
|
||||
|
||||
const ROUND_TRIP_METRIC: &'static str = "surql_trips";
|
||||
// static LOCK: LazyLock<Arc<Mutex<bool>>> = LazyLock::new(|| Arc::new(Mutex::new(true)));
|
||||
static LOCK: LazyLock<Mutex<bool>> = LazyLock::new(|| Mutex::new(true));
|
||||
|
||||
const CUSTOM_ENGINE: engine::GeneralPurpose =
|
||||
engine::GeneralPurpose::new(&alphabet::URL_SAFE, general_purpose::NO_PAD);
|
||||
|
||||
const TIME_SPENT_ON_LOCK: &'static str = "surql_lock_waiting_ms";
|
||||
const STORE: &'static str = "surql_store_calls";
|
||||
const LINK: &'static str = "surql_link_calls";
|
||||
|
||||
#[derive(Serialize, Deserialize, Clone)]
|
||||
#[derive(Deserialize, Clone)]
|
||||
pub struct Website {
|
||||
/// The url that this data is found at
|
||||
pub site: Url,
|
||||
/// Wether or not this link has been crawled yet
|
||||
pub crawled: bool,
|
||||
#[serde(skip_serializing)]
|
||||
id: Option<Thing>,
|
||||
}
|
||||
|
||||
impl Serialize for Website {
|
||||
fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
|
||||
where
|
||||
S: serde::Serializer {
|
||||
let mut state = serializer.serialize_struct("Website", 2)?;
|
||||
state.serialize_field("crawled", &self.crawled)?;
|
||||
// to_string() calls the correct naming of site
|
||||
state.serialize_field("site", &self.site.to_string())?;
|
||||
state.end()
|
||||
}
|
||||
}
|
||||
|
||||
// manual impl to make tracing look nicer
|
||||
impl Debug for Website {
|
||||
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
|
||||
let site = (self.site.domain().unwrap_or("n/a")).to_string() + self.site.path();
|
||||
f.debug_struct("Website").field("site", &site).finish()
|
||||
f.debug_struct("Website").field("site", &self.site).finish()
|
||||
}
|
||||
}
|
||||
|
||||
@ -39,9 +64,8 @@ impl Website {
|
||||
Err(_) => todo!(),
|
||||
};
|
||||
Self {
|
||||
id: None,
|
||||
crawled,
|
||||
site,
|
||||
site
|
||||
}
|
||||
}
|
||||
|
||||
@ -50,24 +74,44 @@ impl Website {
|
||||
self.crawled = true
|
||||
}
|
||||
|
||||
pub fn get_url_as_string(site: &Url) -> String {
|
||||
let domain = match site.domain() {
|
||||
Some(s) => s.to_string(),
|
||||
None => {
|
||||
warn!("Failed to get domain of URL: {}, falling back to 'localhost'", site.to_string());
|
||||
"localhost".to_string()
|
||||
}
|
||||
};
|
||||
let path = site.path();
|
||||
|
||||
domain + path
|
||||
}
|
||||
pub fn get_url_as_b64_path(site: &Url) -> String {
|
||||
let domain = site.domain().unwrap_or("DOMAIN").to_string();
|
||||
let path = &CUSTOM_ENGINE.encode(site.path());
|
||||
|
||||
domain + path
|
||||
}
|
||||
|
||||
#[instrument(skip_all)]
|
||||
pub async fn links_to(&self, other: Vec<Thing>, db: &Surreal<Client>) {
|
||||
|
||||
let len = other.len();
|
||||
if len == 0 {return}
|
||||
if len == 0 {
|
||||
return;
|
||||
}
|
||||
|
||||
let from = &self.site;
|
||||
|
||||
let from = self.site.to_string();
|
||||
// let to = other.site.to_string();
|
||||
trace!("Linking {from} to {} other pages.", other.len());
|
||||
let msg = format!("Linked {len} pages");
|
||||
trace!("Linking {} pages to {from}", other.len());
|
||||
let msg = format!("Linked {len} pages to {from}");
|
||||
let timer = Timer::start(&msg);
|
||||
// prevent the timer from being dropped instantly.
|
||||
let _ = timer;
|
||||
counter!(ROUND_TRIP_METRIC).increment(1);
|
||||
counter!(LINK).increment(1);
|
||||
match db
|
||||
.query("COUNT(RELATE (SELECT id FROM website WHERE site = $in) -> links_to -> $out)")
|
||||
.bind(("in", from))
|
||||
.bind(("in", from.clone()))
|
||||
.bind(("out", other))
|
||||
.await
|
||||
{
|
||||
@ -79,90 +123,64 @@ impl Website {
|
||||
let _: Vec<usize> = vec;
|
||||
if let Some(num) = vec.get(0) {
|
||||
if *num == len {
|
||||
trace!("Link OK");
|
||||
trace!("Link for {from} OK - {num}/{len}");
|
||||
return;
|
||||
} else {
|
||||
warn!("Didn't link all the records. {num}/{len}");
|
||||
error!("Didn't link all the records. {num}/{len}. Surreal response: {:?}", e);
|
||||
return;
|
||||
}
|
||||
}
|
||||
}
|
||||
warn!("Linking request succeeded but couldn't verify the results.");
|
||||
},
|
||||
error!("Linking request succeeded but couldn't verify the results.");
|
||||
}
|
||||
Err(e) => {
|
||||
error!("{}", e.to_string());
|
||||
},
|
||||
}
|
||||
}
|
||||
|
||||
#[instrument(name = "surql_store", skip_all)]
|
||||
pub async fn store(&self, db: &Surreal<Client>) -> Option<Thing> {
|
||||
counter!(STORE).increment(1);
|
||||
let counter = counter!(ROUND_TRIP_METRIC);
|
||||
let t = Timer::start("Stored link");
|
||||
let _ = t;
|
||||
counter.increment(1);
|
||||
|
||||
// check if it's been gone thru before
|
||||
let mut response = db
|
||||
.query("SELECT * FROM ONLY website WHERE site = $site LIMIT 1")
|
||||
.bind(("site", self.site.to_string()))
|
||||
.await
|
||||
.expect("Failed to check surreal for duplicates!");
|
||||
|
||||
if let Some(old) = response.take::<Option<Website>>(0).expect("Failed to read response from surreal for duplicates.") {
|
||||
// site exists already
|
||||
if let Some(id) = old.id {
|
||||
// make sure to preserve the "crawled status"
|
||||
let mut new = self.clone();
|
||||
new.crawled = old.crawled | new.crawled;
|
||||
|
||||
counter.increment(1);
|
||||
// update the record
|
||||
match db.upsert((id.tb, id.id.to_string())).content(new).await {
|
||||
Ok(e) => {
|
||||
if let Some(a) = e {
|
||||
let _: Record = a;
|
||||
return Some(a.id);
|
||||
}
|
||||
}
|
||||
Err(e) => {
|
||||
match e {
|
||||
surrealdb::Error::Db(error) => {
|
||||
match error {
|
||||
Db::QueryCancelled => todo!(),
|
||||
Db::QueryNotExecuted => todo!(),
|
||||
Db::QueryNotExecutedDetail { message: _ } => todo!(),
|
||||
_=>{},
|
||||
}
|
||||
},
|
||||
_=>{},
|
||||
}
|
||||
// error!("{}", e);
|
||||
}
|
||||
};
|
||||
}
|
||||
} else {
|
||||
counter.increment(1);
|
||||
// sites hasn't existed yet
|
||||
match db.create("website").content(self.clone()).await {
|
||||
Ok(e) => {
|
||||
let _: Option<Record> = e;
|
||||
if let Some(a) = e {
|
||||
let _: Record = a;
|
||||
return Some(a.id);
|
||||
}
|
||||
}
|
||||
Err(a) => error!("{:?}", a),
|
||||
};
|
||||
}
|
||||
None
|
||||
}
|
||||
}
|
||||
|
||||
impl ToString for Website {
|
||||
fn to_string(&self) -> String {
|
||||
self.site.to_string()
|
||||
// Insert ever item in the vec into surreal, crawled state will be preserved as TRUE
|
||||
// if already in the database as such or incoming data is TRUE.
|
||||
pub async fn store_all(all: Vec<Self>, db: &Surreal<Client>) -> Vec<Thing> {
|
||||
counter!(STORE).increment(1);
|
||||
let mut things = Vec::with_capacity(all.len());
|
||||
|
||||
// TODO this only allows for one thread to be in the database at a time.
|
||||
// This is currently required since otherwise we get write errors.
|
||||
// If the default `crawled` is set to false, we might not need to write any more
|
||||
// than just the name. `accessed_at` is fun but not needed.
|
||||
let now = Instant::now();
|
||||
let lock = LOCK.lock().await;
|
||||
counter!(TIME_SPENT_ON_LOCK).increment(now.elapsed().as_millis() as u64);
|
||||
|
||||
match db
|
||||
.query(
|
||||
"INSERT INTO website $array
|
||||
ON DUPLICATE KEY UPDATE
|
||||
accessed_at = time::now(),
|
||||
crawled = crawled OR $input.crawled
|
||||
RETURN VALUE id;
|
||||
",
|
||||
)
|
||||
.bind(("array", all))
|
||||
.await
|
||||
{
|
||||
Ok(mut id) => match id.take::<Vec<Thing>>(0) {
|
||||
Ok(mut x) => things.append(&mut x),
|
||||
Err(err) => match err {
|
||||
Api(error) => {
|
||||
eprintln!("{:?}", error);
|
||||
error!("{:?}", error);
|
||||
}
|
||||
_ => error!("{:?}", err),
|
||||
},
|
||||
},
|
||||
Err(err) => {
|
||||
error!("{:?}", err);
|
||||
}
|
||||
}
|
||||
drop(lock);
|
||||
things
|
||||
}
|
||||
}
|
||||
|
||||
@ -193,15 +211,16 @@ pub async fn connect(config: &Config) -> surrealdb::Result<Surreal<Client>> {
|
||||
.await?;
|
||||
|
||||
// Select a specific namespace / database
|
||||
db
|
||||
.use_ns(&config.surreal_ns)
|
||||
db.use_ns(&config.surreal_ns)
|
||||
.use_db(&config.surreal_db)
|
||||
.await?;
|
||||
|
||||
|
||||
let setup = include_bytes!("setup.surql");
|
||||
let file = setup.iter().map(|c| *c as char).collect::<String>();
|
||||
|
||||
db.query(file).await.expect("Failed to setup surreal tables.");
|
||||
db.query(file)
|
||||
.await
|
||||
.expect("Failed to setup surreal tables.");
|
||||
|
||||
Ok(db)
|
||||
}
|
||||
|
95
src/main.rs
95
src/main.rs
@ -2,7 +2,12 @@
|
||||
|
||||
extern crate html5ever;
|
||||
|
||||
use std::{fs::File, io::Read, net::{IpAddr, Ipv4Addr}, time::Instant};
|
||||
use std::{
|
||||
fs::File,
|
||||
io::Read,
|
||||
net::{IpAddr, Ipv4Addr},
|
||||
time::Instant,
|
||||
};
|
||||
|
||||
use db::{connect, Website};
|
||||
use metrics::{counter, gauge};
|
||||
@ -11,7 +16,7 @@ use s3::S3;
|
||||
use serde::Deserialize;
|
||||
use surrealdb::{engine::remote::ws::Client, Surreal};
|
||||
use tokio::task::JoinSet;
|
||||
use tracing::{debug, info, instrument, trace, trace_span, warn};
|
||||
use tracing::{debug, error, info, instrument, trace, trace_span, warn};
|
||||
use tracing_subscriber::{fmt, layer::SubscriberExt, EnvFilter, Layer, Registry};
|
||||
|
||||
mod db;
|
||||
@ -50,25 +55,26 @@ async fn main() {
|
||||
.open("./docker/logs/tracing.log")
|
||||
.expect("Couldn't make log file!");
|
||||
|
||||
let registry = Registry::default()
|
||||
.with(
|
||||
fmt::layer()
|
||||
.with_line_number(true)
|
||||
.with_thread_ids(true)
|
||||
.with_file(true)
|
||||
// .with_timer(LocalTime::rfc_3339()) // Loki or alloy does this automatically
|
||||
.json()
|
||||
.with_writer(writer)
|
||||
// .with_filter(EnvFilter::from_default_env())
|
||||
);
|
||||
let filter = EnvFilter::from_default_env();
|
||||
|
||||
let registry = Registry::default().with(
|
||||
fmt::layer()
|
||||
.with_line_number(true)
|
||||
.with_thread_ids(true)
|
||||
.with_file(true)
|
||||
.json()
|
||||
.with_writer(writer)
|
||||
.with_filter(filter)
|
||||
);
|
||||
|
||||
tracing::subscriber::set_global_default(registry).expect("Failed to set default subscriber");
|
||||
|
||||
|
||||
let builder = PrometheusBuilder::new();
|
||||
builder.with_http_listener(
|
||||
std::net::SocketAddr::new(IpAddr::V4(Ipv4Addr::from_octets([0,0,0,0])), 2500)
|
||||
)
|
||||
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");
|
||||
|
||||
@ -80,7 +86,6 @@ async fn main() {
|
||||
// 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);
|
||||
@ -106,7 +111,7 @@ async fn main() {
|
||||
let pre_loop_span = span.enter();
|
||||
// Download the site
|
||||
let site = Website::new(&starting_url, false);
|
||||
get(site, db.clone(), reqwest.clone(), s3.clone()).await;
|
||||
process(site, db.clone(), reqwest.clone(), s3.clone()).await;
|
||||
|
||||
drop(pre_loop_span);
|
||||
|
||||
@ -133,7 +138,7 @@ async fn main() {
|
||||
let mut futures = JoinSet::new();
|
||||
for site in uncrawled {
|
||||
gauge!(BEING_PROCESSED).increment(1);
|
||||
futures.spawn(get(site, db.clone(), reqwest.clone(), s3.clone()));
|
||||
futures.spawn(process(site, db.clone(), reqwest.clone(), s3.clone()));
|
||||
// let percent = format!("{:.2}%", (crawled as f32 / budget as f32) * 100f32);
|
||||
// info!("Crawled {crawled} out of {budget} pages. ({percent})");
|
||||
}
|
||||
@ -150,40 +155,60 @@ async fn main() {
|
||||
}
|
||||
drop(span);
|
||||
|
||||
info!("Done");
|
||||
debug!("Done");
|
||||
drop(total_runtime);
|
||||
}
|
||||
|
||||
#[instrument(skip (db, s3, reqwest))]
|
||||
#[instrument(skip(db, s3, 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 get(mut site: Website, db: Surreal<Client>, reqwest: reqwest::Client, s3: S3) {
|
||||
trace!("Get: {}", site.to_string());
|
||||
|
||||
/// 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, s3: S3) {
|
||||
|
||||
// METRICS
|
||||
trace!("Process: {}", &site.site);
|
||||
let timer = Timer::start("Built request");
|
||||
let request_builder = reqwest.get(site.to_string());
|
||||
// Build the request
|
||||
let request_builder = reqwest.get(&site.site.to_string());
|
||||
// METRICS
|
||||
timer.stop();
|
||||
|
||||
// METRICS
|
||||
let g = gauge!(GET_IN_FLIGHT);
|
||||
g.increment(1);
|
||||
let timer = Timer::start("Got page");
|
||||
|
||||
// Send the http request (get)
|
||||
if let Ok(response) = request_builder.send().await {
|
||||
|
||||
// METRICS
|
||||
timer.stop();
|
||||
g.decrement(1);
|
||||
counter!(GET_METRIC).increment(1);
|
||||
debug!("Getting body...");
|
||||
|
||||
// Get body
|
||||
let data = response.text().await.expect("Failed to read http response's body!");
|
||||
// Get body from response
|
||||
let data = response
|
||||
.text()
|
||||
.await
|
||||
.expect("Failed to read http response's body!");
|
||||
// Store document
|
||||
s3.store(&data, &site.site).await;
|
||||
// Parse document and store relationships
|
||||
parser::parse(&db, &mut site, &data).await;
|
||||
return;
|
||||
|
||||
// Parse document and get relationships
|
||||
let sites = parser::parse(&site, &data).await;
|
||||
|
||||
// update self in db
|
||||
site.set_crawled();
|
||||
Website::store_all(vec![site.clone()], &db).await;
|
||||
|
||||
// Store all the other sites so that we can link to them.
|
||||
// let mut links_to = Vec::new();
|
||||
let others = Website::store_all(sites, &db).await;
|
||||
|
||||
// Make the database's links reflect the html links between sites
|
||||
site.links_to(others, &db).await;
|
||||
} else {
|
||||
error!("Failed to get: {}", &site.site);
|
||||
}
|
||||
trace!("Failed to get: {}", site.to_string());
|
||||
}
|
||||
|
||||
/// Returns uncrawled links
|
||||
@ -228,8 +253,6 @@ impl<'a> Timer<'a> {
|
||||
|
||||
if ms > 200. {
|
||||
warn!("{}", format!("{} in {:.3}ms", self.msg, ms));
|
||||
} else {
|
||||
trace!("{}", format!("{} in {:.3}ms", self.msg, ms));
|
||||
}
|
||||
|
||||
ms
|
||||
|
@ -5,8 +5,6 @@ use html5ever::tokenizer::{BufferQueue, TokenizerResult};
|
||||
use html5ever::tokenizer::{StartTag, TagToken};
|
||||
use html5ever::tokenizer::{Token, TokenSink, TokenSinkResult, Tokenizer, TokenizerOpts};
|
||||
use html5ever::{local_name, tendril::*};
|
||||
use surrealdb::engine::remote::ws::Client;
|
||||
use surrealdb::Surreal;
|
||||
use tracing::instrument;
|
||||
|
||||
use crate::db::Website;
|
||||
@ -40,7 +38,9 @@ impl TokenSink for Website {
|
||||
// Set url
|
||||
let mut url = web.site;
|
||||
url.set_fragment(None); // removes #xyz
|
||||
let joined = url.join(&attr.value).expect("Failed to join url during parsing!");
|
||||
let joined = url
|
||||
.join(&attr.value)
|
||||
.expect("Failed to join url during parsing!");
|
||||
web.site = joined;
|
||||
|
||||
web.crawled = false;
|
||||
@ -65,45 +65,31 @@ impl TokenSink for Website {
|
||||
}
|
||||
|
||||
#[instrument(skip_all)]
|
||||
pub async fn parse(db: &Surreal<Client>, site: &mut Website, data: &str) {
|
||||
// update self in db
|
||||
site.set_crawled();
|
||||
site.store(db).await;
|
||||
|
||||
/// Parses the passed site and returns all the sites it links to.
|
||||
pub async fn parse(site: &Website, data: &str) -> Vec<Website> {
|
||||
// prep work
|
||||
let mut other_sites: Vec<Website> = Vec::new();
|
||||
{ // using blocks to prevent compiler's async worries
|
||||
let _t = Timer::start("Parsed page");
|
||||
let _t = Timer::start("Parsed page");
|
||||
|
||||
// change data into something that can be tokenized
|
||||
let chunk = Tendril::from_str(&data).expect("Failed to parse string into Tendril!");
|
||||
// create buffer of tokens and push our input into it
|
||||
let mut token_buffer = BufferQueue::default();
|
||||
token_buffer.push_back(chunk.try_reinterpret::<fmt::UTF8>().expect("Failed to reinterprt chunk!"));
|
||||
// create the tokenizer
|
||||
let tokenizer = Tokenizer::new(site.clone(), TokenizerOpts::default());
|
||||
// change data into something that can be tokenized
|
||||
let chunk = Tendril::from_str(&data).expect("Failed to parse string into Tendril!");
|
||||
// create buffer of tokens and push our input into it
|
||||
let mut token_buffer = BufferQueue::default();
|
||||
token_buffer.push_back(
|
||||
chunk
|
||||
.try_reinterpret::<fmt::UTF8>()
|
||||
.expect("Failed to reinterprt chunk!"),
|
||||
);
|
||||
// create the tokenizer
|
||||
let tokenizer = Tokenizer::new(site.clone(), TokenizerOpts::default());
|
||||
|
||||
// go thru buffer
|
||||
while let TokenizerResult::Script(mut sites) = tokenizer.feed(&mut token_buffer) {
|
||||
other_sites.append(&mut sites);
|
||||
// other_sites.push(sites);
|
||||
}
|
||||
|
||||
assert!(token_buffer.is_empty());
|
||||
tokenizer.end();
|
||||
// go thru buffer
|
||||
while let TokenizerResult::Script(mut sites) = tokenizer.feed(&mut token_buffer) {
|
||||
other_sites.append(&mut sites);
|
||||
// other_sites.push(sites);
|
||||
}
|
||||
assert!(token_buffer.is_empty());
|
||||
tokenizer.end();
|
||||
|
||||
{
|
||||
let mut links_to = Vec::with_capacity(other_sites.len());
|
||||
|
||||
for a in other_sites {
|
||||
|
||||
let other = a.store(db).await;
|
||||
if let Some(o) = other {
|
||||
links_to.push(o);
|
||||
}
|
||||
}
|
||||
|
||||
site.links_to(links_to, db).await;
|
||||
}
|
||||
other_sites
|
||||
}
|
||||
|
68
src/s3.rs
68
src/s3.rs
@ -1,4 +1,3 @@
|
||||
use base64::{alphabet, engine::{self, general_purpose}, Engine};
|
||||
use metrics::counter;
|
||||
use minio::s3::{
|
||||
args::{BucketExistsArgs, MakeBucketArgs},
|
||||
@ -11,11 +10,9 @@ use minio::s3::{
|
||||
use tracing::{instrument, trace, warn};
|
||||
use url::Url;
|
||||
|
||||
use crate::{Config, Timer};
|
||||
|
||||
const CUSTOM_ENGINE: engine::GeneralPurpose = engine::GeneralPurpose::new(&alphabet::URL_SAFE, general_purpose::NO_PAD);
|
||||
use crate::{db::Website, Config, Timer};
|
||||
|
||||
const ROUND_TRIP_METRIC: &'static str = "s3_trips";
|
||||
const S3_ROUND_TRIP_METRIC: &'static str = "s3_trips";
|
||||
|
||||
#[derive(Clone)]
|
||||
pub struct S3 {
|
||||
@ -45,6 +42,7 @@ impl S3 {
|
||||
.expect("Failed to check if bucket exists"),
|
||||
)
|
||||
.await?;
|
||||
counter!(S3_ROUND_TRIP_METRIC).increment(1);
|
||||
|
||||
if !exists {
|
||||
trace!("Creating bucket...");
|
||||
@ -54,52 +52,50 @@ impl S3 {
|
||||
)
|
||||
.await?;
|
||||
}
|
||||
counter!(S3_ROUND_TRIP_METRIC).increment(1);
|
||||
|
||||
trace!("Connection successful");
|
||||
|
||||
Ok(Self {
|
||||
bucket_name: config.s3_bucket.to_owned(),
|
||||
client: client,
|
||||
client,
|
||||
})
|
||||
}
|
||||
|
||||
#[instrument(name = "s3_store", skip_all)]
|
||||
pub async fn store(&self, data: &str, url: &Url) {
|
||||
let counter = counter!(ROUND_TRIP_METRIC);
|
||||
let counter = counter!(S3_ROUND_TRIP_METRIC);
|
||||
let t = Timer::start("Stored page");
|
||||
let _ = t; // prevent compiler drop
|
||||
if let Some(domain) = url.domain() {
|
||||
let filename = domain.to_owned() + url.path();
|
||||
|
||||
trace!("Created filename: {filename} from raw: {}", url.to_string());
|
||||
let filename = Website::get_url_as_string(url);
|
||||
trace!("Storing {} as {filename}", url.to_string());
|
||||
|
||||
counter.increment(1);
|
||||
let _ = match &self
|
||||
.client
|
||||
.put_object_content(&self.bucket_name, &filename, data.to_owned())
|
||||
.send()
|
||||
.await {
|
||||
Ok(_) => {},
|
||||
Err(err) => {
|
||||
match err {
|
||||
Error::InvalidObjectName(_) => {
|
||||
counter.increment(1);
|
||||
let _ = match &self
|
||||
.client
|
||||
.put_object_content(&self.bucket_name, &filename, data.to_owned())
|
||||
.send()
|
||||
.await
|
||||
{
|
||||
Ok(_) => {}
|
||||
Err(err) => match err {
|
||||
Error::InvalidObjectName(_) => {
|
||||
// This code will really only run if the url has non-english chars
|
||||
warn!("Tried storing invalid object name, retrying with Base64 encoding. Last try.");
|
||||
|
||||
warn!("Tried storing invalid object name, retrying with Base64 encoding. Last try.");
|
||||
|
||||
let filename: String = domain.to_owned() + &CUSTOM_ENGINE.encode(url.path());
|
||||
let filename: String = Website::get_url_as_b64_path(url);
|
||||
|
||||
counter.increment(1);
|
||||
let _ = &self
|
||||
.client
|
||||
.put_object_content(&self.bucket_name, &filename, data.to_owned())
|
||||
.send()
|
||||
.await
|
||||
.unwrap();
|
||||
},
|
||||
_ => {},
|
||||
}
|
||||
},
|
||||
};
|
||||
}
|
||||
counter.increment(1);
|
||||
let _ = &self
|
||||
.client
|
||||
.put_object_content(&self.bucket_name, &filename, data.to_owned())
|
||||
.send()
|
||||
.await
|
||||
.unwrap();
|
||||
}
|
||||
_ => {}
|
||||
},
|
||||
};
|
||||
}
|
||||
}
|
||||
|
@ -1,3 +1,9 @@
|
||||
DEFINE TABLE IF NOT EXISTS website SCHEMALESS;
|
||||
DEFINE FIELD IF NOT EXISTS accessed_at ON TABLE website VALUE time::now();
|
||||
|
||||
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 accessed_at ON TABLE website VALUE time::now();
|
||||
DEFINE FIELD IF NOT EXISTS first_accessed_at ON TABLE website VALUE time::now();
|
||||
|
Loading…
x
Reference in New Issue
Block a user