hugo/parser/metadecoders/decoder.go
Bjørn Erik Pedersen 7285e74090
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning
There are some breaking changes in this commit, see #11455.

Closes #11455
Closes #11549

This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build.

The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate.

A list of the notable new features:

* A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server.
* A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently.
* You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs.
We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections).
Memory Limit
* Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory.

New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively.

This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive):

Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get
Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers.
Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds.

Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role).

Fixes #10169
Fixes #10364
Fixes #10482
Fixes #10630
Fixes #10656
Fixes #10694
Fixes #10918
Fixes #11262
Fixes #11439
Fixes #11453
Fixes #11457
Fixes #11466
Fixes #11540
Fixes #11551
Fixes #11556
Fixes #11654
Fixes #11661
Fixes #11663
Fixes #11664
Fixes #11669
Fixes #11671
Fixes #11807
Fixes #11808
Fixes #11809
Fixes #11815
Fixes #11840
Fixes #11853
Fixes #11860
Fixes #11883
Fixes #11904
Fixes #7388
Fixes #7425
Fixes #7436
Fixes #7544
Fixes #7882
Fixes #7960
Fixes #8255
Fixes #8307
Fixes #8863
Fixes #8927
Fixes #9192
Fixes #9324
2024-01-27 16:28:14 +01:00

318 lines
7.8 KiB
Go

// Copyright 2018 The Hugo Authors. All rights reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package metadecoders
import (
"bytes"
"encoding/csv"
"encoding/json"
"fmt"
"log"
"regexp"
"strconv"
"strings"
"github.com/gohugoio/hugo/common/herrors"
"github.com/gohugoio/hugo/common/maps"
"github.com/niklasfasching/go-org/org"
xml "github.com/clbanning/mxj/v2"
toml "github.com/pelletier/go-toml/v2"
"github.com/spf13/afero"
"github.com/spf13/cast"
yaml "gopkg.in/yaml.v2"
)
// Decoder provides some configuration options for the decoders.
type Decoder struct {
// Delimiter is the field delimiter used in the CSV decoder. It defaults to ','.
Delimiter rune
// Comment, if not 0, is the comment character used in the CSV decoder. Lines beginning with the
// Comment character without preceding whitespace are ignored.
Comment rune
// If true, a quote may appear in an unquoted field and a non-doubled quote
// may appear in a quoted field. It defaults to false.
LazyQuotes bool
}
// OptionsKey is used in cache keys.
func (d Decoder) OptionsKey() string {
var sb strings.Builder
sb.WriteRune(d.Delimiter)
sb.WriteRune(d.Comment)
sb.WriteString(strconv.FormatBool(d.LazyQuotes))
return sb.String()
}
// Default is a Decoder in its default configuration.
var Default = Decoder{
Delimiter: ',',
}
// UnmarshalToMap will unmarshall data in format f into a new map. This is
// what's needed for Hugo's front matter decoding.
func (d Decoder) UnmarshalToMap(data []byte, f Format) (map[string]any, error) {
m := make(map[string]any)
if data == nil {
return m, nil
}
err := d.UnmarshalTo(data, f, &m)
return m, err
}
// UnmarshalFileToMap is the same as UnmarshalToMap, but reads the data from
// the given filename.
func (d Decoder) UnmarshalFileToMap(fs afero.Fs, filename string) (map[string]any, error) {
format := FormatFromString(filename)
if format == "" {
return nil, fmt.Errorf("%q is not a valid configuration format", filename)
}
data, err := afero.ReadFile(fs, filename)
if err != nil {
return nil, err
}
return d.UnmarshalToMap(data, format)
}
// UnmarshalStringTo tries to unmarshal data to a new instance of type typ.
func (d Decoder) UnmarshalStringTo(data string, typ any) (any, error) {
data = strings.TrimSpace(data)
// We only check for the possible types in YAML, JSON and TOML.
switch typ.(type) {
case string:
return data, nil
case map[string]any, maps.Params:
format := d.FormatFromContentString(data)
return d.UnmarshalToMap([]byte(data), format)
case []any:
// A standalone slice. Let YAML handle it.
return d.Unmarshal([]byte(data), YAML)
case bool:
return cast.ToBoolE(data)
case int:
return cast.ToIntE(data)
case int64:
return cast.ToInt64E(data)
case float64:
return cast.ToFloat64E(data)
default:
return nil, fmt.Errorf("unmarshal: %T not supported", typ)
}
}
// Unmarshal will unmarshall data in format f into an interface{}.
// This is what's needed for Hugo's /data handling.
func (d Decoder) Unmarshal(data []byte, f Format) (any, error) {
if len(data) == 0 {
switch f {
case CSV:
return make([][]string, 0), nil
default:
return make(map[string]any), nil
}
}
var v any
err := d.UnmarshalTo(data, f, &v)
return v, err
}
// UnmarshalTo unmarshals data in format f into v.
func (d Decoder) UnmarshalTo(data []byte, f Format, v any) error {
var err error
switch f {
case ORG:
err = d.unmarshalORG(data, v)
case JSON:
err = json.Unmarshal(data, v)
case XML:
var xmlRoot xml.Map
xmlRoot, err = xml.NewMapXml(data)
var xmlValue map[string]any
if err == nil {
xmlRootName, err := xmlRoot.Root()
if err != nil {
return toFileError(f, data, fmt.Errorf("failed to unmarshal XML: %w", err))
}
xmlValue = xmlRoot[xmlRootName].(map[string]any)
}
switch v := v.(type) {
case *map[string]any:
*v = xmlValue
case *any:
*v = xmlValue
}
case TOML:
err = toml.Unmarshal(data, v)
case YAML:
err = yaml.Unmarshal(data, v)
if err != nil {
return toFileError(f, data, fmt.Errorf("failed to unmarshal YAML: %w", err))
}
// To support boolean keys, the YAML package unmarshals maps to
// map[interface{}]interface{}. Here we recurse through the result
// and change all maps to map[string]interface{} like we would've
// gotten from `json`.
var ptr any
switch vv := v.(type) {
case *map[string]any:
ptr = *vv
case *any:
ptr = *vv
default:
// Not a map.
}
if ptr != nil {
if mm, changed := stringifyMapKeys(ptr); changed {
switch vv := v.(type) {
case *map[string]any:
*vv = mm.(map[string]any)
case *any:
*vv = mm
}
}
}
case CSV:
return d.unmarshalCSV(data, v)
default:
return fmt.Errorf("unmarshal of format %q is not supported", f)
}
if err == nil {
return nil
}
return toFileError(f, data, fmt.Errorf("unmarshal failed: %w", err))
}
func (d Decoder) unmarshalCSV(data []byte, v any) error {
r := csv.NewReader(bytes.NewReader(data))
r.Comma = d.Delimiter
r.Comment = d.Comment
r.LazyQuotes = d.LazyQuotes
records, err := r.ReadAll()
if err != nil {
return err
}
switch vv := v.(type) {
case *any:
*vv = records
default:
return fmt.Errorf("CSV cannot be unmarshaled into %T", v)
}
return nil
}
func parseORGDate(s string) string {
r := regexp.MustCompile(`[<\[](\d{4}-\d{2}-\d{2}) .*[>\]]`)
if m := r.FindStringSubmatch(s); m != nil {
return m[1]
}
return s
}
func (d Decoder) unmarshalORG(data []byte, v any) error {
config := org.New()
config.Log = log.Default() // TODO(bep)
document := config.Parse(bytes.NewReader(data), "")
if document.Error != nil {
return document.Error
}
frontMatter := make(map[string]any, len(document.BufferSettings))
for k, v := range document.BufferSettings {
k = strings.ToLower(k)
if strings.HasSuffix(k, "[]") {
frontMatter[k[:len(k)-2]] = strings.Fields(v)
} else if strings.Contains(v, "\n") {
frontMatter[k] = strings.Split(v, "\n")
} else if k == "date" || k == "lastmod" || k == "publishdate" || k == "expirydate" {
frontMatter[k] = parseORGDate(v)
} else {
frontMatter[k] = v
}
}
switch vv := v.(type) {
case *map[string]any:
*vv = frontMatter
case *any:
*vv = frontMatter
}
return nil
}
func toFileError(f Format, data []byte, err error) error {
return herrors.NewFileErrorFromName(err, fmt.Sprintf("_stream.%s", f)).UpdateContent(bytes.NewReader(data), nil)
}
// stringifyMapKeys recurses into in and changes all instances of
// map[interface{}]interface{} to map[string]interface{}. This is useful to
// work around the impedance mismatch between JSON and YAML unmarshaling that's
// described here: https://github.com/go-yaml/yaml/issues/139
//
// Inspired by https://github.com/stripe/stripe-mock, MIT licensed
func stringifyMapKeys(in any) (any, bool) {
switch in := in.(type) {
case []any:
for i, v := range in {
if vv, replaced := stringifyMapKeys(v); replaced {
in[i] = vv
}
}
case map[string]any:
for k, v := range in {
if vv, changed := stringifyMapKeys(v); changed {
in[k] = vv
}
}
case map[any]any:
res := make(map[string]any)
var (
ok bool
err error
)
for k, v := range in {
var ks string
if ks, ok = k.(string); !ok {
ks, err = cast.ToStringE(k)
if err != nil {
ks = fmt.Sprintf("%v", k)
}
}
if vv, replaced := stringifyMapKeys(v); replaced {
res[ks] = vv
} else {
res[ks] = v
}
}
return res, true
}
return nil, false
}