Cleaning things up to turn this into a proper parser

dev
Nick Dumas 2 years ago
parent 0b2443c720
commit c3b5d8e663

4
.gitignore vendored

@ -0,0 +1,4 @@
node_modules/*
package*.json
dist/*
reports/*

@ -0,0 +1,217 @@
# This file is intended as a starting point for a customized makefile for a Go project.
#
# Targets:
# all: Format, check, build, and test the code
# setup: Install build/test toolchain dependencies (e.g. gox)
# lint: Run linters against source code
# bump-{major,minor,patch}: create a new semver git tag
# release-{major,minor,patch}: push a tagged release
# format: Format the source files
# build: Build the command(s) for target OS/arch combinations
# install: Install the command(s)
# clean: Clean the build/test artifacts
# report: Generate build/test reports
# check: Run tests
# bench: Run benchmarks
# dist: zip/tar binaries & documentation
# debug: print parameters
#
# Parameters:
# VERSION: release version in semver format
# BUILD_TAGS: additional build tags to pass to go build
# DISTDIR: path to save distribution files
# RPTDIR: path to save build/test reports
#
# Assumptions:
# - Your package contains a cmd/ package, containing a directory for each produced binary.
# - You have cloc installed and accessible in the PATH.
# - Your GOPATH and GOROOT are set correctly.
# - Your makefile is in the root of your package and does not have a space in its file name.
# - Your root package contains global string variables Version and Build, to receive the bild version number and commit ID, respectively.
#
# Features:
# - report generates files that can be consumed by Jenkins, as well as a list of external dependencies.
# - setup installs all the tools aside from cloc.
# - Works on Windows and with paths containing spaces.
# - Works when executing from outside the makefile directory using -f.
# - Targets are useful both in CI and developer workstations.
# - Handles cross-compiation for multiple OSes and architectures.
# - Bundles binaries and documentation into compressed archives, using tar/gz for Linux and Darwin, and zip for Windows.
# Parameters
PKG = code.ndumas.com/ndumas/wikilink-parser
NAME = parse-wikilinks
DOC = README.md LICENSE
# Replace backslashes with forward slashes for use on Windows.
# Make is !@#$ing weird.
E :=
BSLASH := \$E
FSLASH := /
# Directories
WD := $(subst $(BSLASH),$(FSLASH),$(shell pwd))
MD := $(subst $(BSLASH),$(FSLASH),$(shell dirname "$(realpath $(lastword $(MAKEFILE_LIST)))"))
PKGDIR = $(MD)
CMDDIR = $(PKGDIR)/cmd
DISTDIR ?= $(WD)/dist
RPTDIR ?= $(WD)/reports
GP = $(subst $(BSLASH),$(FSLASH),$(GOPATH))
# Parameters
VERSION ?= $(shell git -C "$(MD)" describe --tags --dirty=-dev)
COMMIT_ID := $(shell git -C "$(MD)" rev-parse HEAD | head -c8)
BUILD_TAGS ?= release
CMDPKG = $(PKG)/cmd
CMDS := $(shell find "$(CMDDIR)/" -mindepth 1 -maxdepth 1 -type d | sed 's/ /\\ /g' | xargs -n1 basename)
BENCHCPUS ?= 1,2,4
# Commands
GOCMD = go
ARCHES ?= amd64 386
OSES ?= windows linux darwin
OUTTPL = $(DISTDIR)/$(NAME)-$(VERSION)-{{.OS}}_{{.Arch}}/{{.Dir}}
LDFLAGS = -X $(PKG).Version=$(VERSION) -X $(PKG).Build=$(COMMIT_ID)
GOBUILD = gox -osarch="!darwin/386" -rebuild -gocmd="$(GOCMD)" -arch="$(ARCHES)" -os="$(OSES)" -output="$(OUTTPL)" -tags "$(BUILD_TAGS)" -ldflags "$(LDFLAGS)"
GOCLEAN = $(GOCMD) clean
GOINSTALL = $(GOCMD) install -a -tags "$(BUILD_TAGS)" -ldflags "$(LDFLAGS)"
GOTEST = $(GOCMD) test -v -tags "$(BUILD_TAGS)"
DISABLED_LINTERS = varnamelen,interfacer,ifshort,exhaustivestruct,maligned,varcheck,scopelint,structcheck,deadcode,nosnakecase,golint,depguard
GOLINT = golangci-lint run --enable-all --disable "$(DISABLED_LINTERS)" --timeout=30s --tests
GODEP = $(GOCMD) get -d -t
GOFMT = goreturns -w
GOBENCH = $(GOCMD) test -v -tags "$(BUILD_TAGS)" -cpu=$(BENCHCPUS) -run=NOTHING -bench=. -benchmem -outputdir "$(RPTDIR)"
GZCMD = tar -czf
ZIPCMD = zip
SHACMD = sha256sum
SLOCCMD = cloc --by-file --xml --exclude-dir="vendor" --include-lang="Go"
XUCMD = go2xunit
DOCKER_CMD=docker --config ~/.docker/
# Dynamic Targets
INSTALL_TARGETS := $(addprefix install-,$(CMDS))
.PHONY: all
all: debug setup dep format lint test bench build dist
git-push:
git push origin main --tags
git push github main --tags
release-major: bump-major git-push
release-minor: bump-minor git-push
release-patch: bump-patch git-push
setup: setup-dirs setup-build setup-format setup-lint setup-reports setup-bump
setup-bump:
go install github.com/guilhem/bump@latest
bump-major: setup-bump
bump major
bump-minor: setup-bump
bump minor
bump-patch: setup-bump
bump patch
setup-reports: setup-dirs
go install github.com/tebeka/go2xunit@latest
setup-build: setup-dirs
go install github.com/mitchellh/gox@latest
setup-format: setup-dirs
go install github.com/sqs/goreturns@latest
setup-lint: setup-dirs
go install github.com/golangci/golangci-lint/cmd/golangci-lint@v1.53.1
setup-dirs:
mkdir -p "$(RPTDIR)"
mkdir -p "$(DISTDIR)"
clean:
$(GOCLEAN) $(PKG)
rm -vrf "$(DISTDIR)"/*
rm -vf "$(RPTDIR)"/*
format:
$(GOFMT) "$(PKGDIR)"
dep:
$(GODEP) $(PKG)/...
lint: setup-dirs dep
$(GOLINT) "$(PKGDIR)" | tee "$(RPTDIR)/lint.out"
check: setup-dirs clean dep
$(GOTEST) $$(go list "$(PKG)/..." | grep -v /vendor/) | tee "$(RPTDIR)/test.out"
bench: setup-dirs clean dep
$(GOBENCH) $$(go list "$(PKG)/..." | grep -v /vendor/) | tee "$(RPTDIR)/bench.out"
report: check
cd "$(PKGDIR)";$(SLOCCMD) --out="$(RPTDIR)/cloc.xml" . | tee "$(RPTDIR)/cloc.out"
cat "$(RPTDIR)/test.out" | $(XUCMD) -output "$(RPTDIR)/tests.xml"
go list -f '{{join .Deps "\n"}}' "$(CMDPKG)/..." | sort | uniq | xargs -I {} sh -c "go list -f '{{if not .Standard}}{{.ImportPath}}{{end}}' {} | tee -a '$(RPTDIR)/deps.out'"
build: $(CMDS)
$(CMDS): setup-dirs dep
$(GOBUILD) "$(CMDPKG)/$@" | tee "$(RPTDIR)/build-$@.out"
install: $(INSTALL_TARGETS)
$(INSTALL_TARGETS):
$(GOINSTALL) "$(CMDPKG)/$(subst install-,,$@)"
dist: clean build
for docfile in $(DOC); do \
for dir in "$(DISTDIR)"/*; do \
cp "$(PKGDIR)/$$docfile" "$$dir/"; \
done; \
done
cd "$(DISTDIR)"; for dir in ./*linux*; do $(GZCMD) "$(basename "$$dir").tar.gz" "$$dir"; done
cd "$(DISTDIR)"; for dir in ./*windows*; do $(ZIPCMD) "$(basename "$$dir").zip" "$$dir"; done
cd "$(DISTDIR)"; for dir in ./*darwin*; do $(GZCMD) "$(basename "$$dir").tar.gz" "$$dir"; done
cd "$(DISTDIR)"; find . -maxdepth 1 -type f -printf "$(SHACMD) %P | tee \"./%P.sha\"\n" | sh
$(info "Built v$(VERSION), build $(COMMIT_ID)")
.PHONY: docker
docker: docker-image docker-push
.PHONY: docker-push
docker-push:
$(DOCKER_CMD) tag code.ndumas.com/ndumas/wikilink-parser:$(VERSION) code.ndumas.com/ndumas/wikilink-parser:latest
$(DOCKER_CMD) push code.ndumas.com/ndumas/wikilink-parser:latest
$(DOCKER_CMD) push code.ndumas.com/ndumas/wikilink-parser:$(VERSION)
.PHONY: docker-image
docker-image:
$(DOCKER_CMD) build --build-arg VERSION=$(VERSION) -t code.ndumas.com/ndumas/wikilink-parser:$(VERSION) .
.PHONY: build-alpine
build-alpine:
# this version breaks build variable injection
# CGO_ENABLED=0 GOOS=linux go build -ldflags="buildmode=exe $(LDFLAGS) -linkmode external -w -extldflags '-static' " -o $(DISTDIR)/$(NAME)-$(VERSION)-alpine/obp cmd/obp/*.go
CGO_ENABLED=0 GOOS=linux go build -ldflags="$(LDFLAGS)" -o $(DISTDIR)/$(NAME)-$(VERSION)-alpine/obp cmd/obp/*.go
debug:
$(info MD=$(MD))
$(info WD=$(WD))
$(info PKG=$(PKG))
$(info PKGDIR=$(PKGDIR))
$(info DISTDIR=$(DISTDIR))
$(info VERSION=$(VERSION))
$(info COMMIT_ID=$(COMMIT_ID))
$(info BUILD_TAGS=$(BUILD_TAGS))
$(info CMDS=$(CMDS))
$(info BUILD_TARGETS=$(BUILD_TARGETS))
$(info INSTALL_TARGETS=$(INSTALL_TARGETS))

@ -1,3 +1,3 @@
module wikilinks
module code.ndumas.com/ndumas/wikilink-parser
go 1.19

@ -0,0 +1 @@
package wikilink

@ -1,54 +0,0 @@
package main
import (
"log"
"regexp"
)
var (
alias = regexp.MustCompile(`\|(.*)`)
link = regexp.MustCompile(`^([a-zA-Z ]+)?[\|\#\^\\]?`)
subsection = regexp.MustCompile(`#([\w\s]+)`)
block = regexp.MustCompile(`#\^([a-zA-Z ]+)`)
wholeLink = regexp.MustCompile(`\[\[(.*)\]\]`)
)
type Link struct {
Dest, Alias, Subsection, Block string
}
func ExtractLink(raw string) Link {
var li Link
rawLinkMatch := wholeLink.FindAllStringSubmatch(raw, -1)
var rawLink string
if len(rawLinkMatch) > 0 {
rawLink = rawLinkMatch[0][1]
}
l := link.FindAllStringSubmatch(rawLink, -1)
if len(l) > 0 {
li.Dest = l[0][1]
}
a := alias.FindAllStringSubmatch(rawLink, -1)
if len(a) > 0 {
li.Alias = a[0][1]
}
s := subsection.FindAllStringSubmatch(rawLink, -1)
if len(s) > 0 {
li.Subsection = s[0][1]
}
b := block.FindAllStringSubmatch(rawLink, -1)
if len(b) > 0 {
li.Block = b[0][1]
}
return li
}
func main() {
log.Printf("%#v\n", ExtractLink(`Embedding a link in a bigger block of text [[Regular Link#^link to block]] shouldn't cause any problems `))
}

@ -1,34 +0,0 @@
package main
import (
"testing"
)
func Test_ExtractLink(t *testing.T) {
tcs := []struct {
name string
in string
expected Link
}{
{"link", `[[Regular Link]]`, Link{Dest: "Regular Link"}},
{"transclude+link", `![[Transcluded Link]]`, Link{Dest: "Transcluded Link"}},
{"link+alias", `[[Regular Link|Alias]]`, Link{Dest: "Regular Link", Alias: "Alias"}},
{"link+subsection", `[[Regular Link#Subsection of page]]`, Link{Dest: "Regular Link", Subsection: "Subsection of page"}},
{"link+block", `[[Regular Link#^link to block]]`, Link{Dest: "Regular Link", Block: "link to block"}},
{"link+subsection+alias", `[[Regular Link#Subsection of page|Alias]]`, Link{Dest: "Regular Link", Subsection: "Subsection of page", Alias: "Alias"}},
{"link+block+alias", `[[Regular Link#^link to block|Alias]]`, Link{Dest: "Regular Link", Block: "link to block", Alias: "Alias"}},
{"link+alias+escape", `[[Regular Link\|Alias]]`, Link{Dest: "Regular Link", Alias: "Alias"}},
{"link+subsection+alias+escape", `[[Regular Link#Subsection of page\|Alias]]`, Link{Dest: "Regular Link", Subsection: "Subsection of page", Alias: "Alias"}},
{"link+block+alias+escape", `[[Regular Link#^link to block\|Alias]]`, Link{Dest: "Regular Link", Block: "link to block", Alias: "Alias"}},
}
for _, tc := range tcs {
t.Run(tc.name, func(t *testing.T) {
out := ExtractLink(tc.in)
if out != tc.expected {
t.Logf("got %#v\n", out)
t.Fail()
}
})
}
}

@ -0,0 +1,18 @@
package wikilink
import (
// "log"
// "github.com/13rac1/goldmark-wikilink"
)
type Wikilink struct {
Link string
Fragment string
Alias string
}
func Extract(raw string) Wikilink {
var wl Wikilink
return wl
}

@ -0,0 +1,52 @@
package wikilink_test
import (
"testing"
"code.ndumas.com/ndumas/wikilink-parser"
)
func Test_Extract_Link(t *testing.T) {
t.Parallel()
tcs := []struct {
name string
in string
expected string
}{
{name: "wikilink", in: "[[wikilink]]", expected: "wikilink"},
{name: "wikilink|display name", in: "[[wikilink|display name]]", expected: "wikilink"},
{name: "wikilink|display name|second pipe", in: "[[wikilink|display name|second pipe]]", expected: "wikilink"},
{name: "wikilink with numeric alias|420|second pipe", in: "[[wikilink|420|second pipe]]", expected: "wikilink with numeric aliases"},
{name: "wikilink with spaces in filename", in: "[[wikilink spaces]]", expected: "wikilink"},
{name: "#heading", in: "[[#heading]]", expected: ""},
{name: "wikilink#heading", in: "[[wikilink#heading]]", expected: "wikilink"},
{name: "wikilink#heading|display name", in: "[[wikilink#heading|display name]]", expected: "wikilink"},
{name: "wikilink#heading|display name", in: "[[wikilink#heading|display name]]", expected: "wikilink"},
{name: "wikilink with numeric aliases#heading|420|display name", in: "[[wikilink#heading|420|second pipe]]", expected: "wikilink"},
}
/*
"[[#^blockRef]]"
"[[wikilink#^blockRef]]"
"[[wikilink#^blockRef|display name]]"
"[[wikilink#^blockRef|display name|secondpipe]]"
"[[#^blockRef]]"
"[[wikilink#^blockRef]]"
"[[wikilink#^blockRef|display name|secondpipe]]"
*/
for _, tc := range tcs {
tc := tc
t.Run(tc.name, func(t *testing.T) {
t.Parallel()
out := wikilink.Extract(tc.in)
if out.Link != tc.expected {
t.Logf("got %#v\n", out)
t.Fail()
}
})
}
}
Loading…
Cancel
Save