Merge pull request #355 from okp4/feat/uri_encoded

🧠 Logic: 🔗 implement `uri_encoded/3`

x/logic/interpreter/registry.go

@@ -112,6 +112,7 @@ var registry = map[string]any{
 	"bech32_address/2":          predicate.Bech32Address,
 	"source_file/1":             predicate.SourceFile,
 	"json_prolog/2":             predicate.JSONProlog,
+	"uri_encoded/3":             predicate.URIEncoded,
 }
 
 // RegistryNames is the list of the predicate names in the Registry.

x/logic/predicate/uri.go

@@ -0,0 +1,171 @@
+package predicate
+
+import (
+	"context"
+	"fmt"
+	"net/url"
+
+	"github.com/ichiban/prolog/engine"
+	"github.com/okp4/okp4d/x/logic/util"
+)
+
+type Component string
+
+const (
+	QueryComponent    Component = "query"
+	FragmentComponent Component = "fragment"
+	PathComponent     Component = "path"
+	SegmentComponent  Component = "segment"
+)
+
+const upperhex = "0123456789ABCDEF"
+
+func NewComponent(v string) (Component, error) {
+	switch v {
+	case string(QueryComponent):
+		return QueryComponent, nil
+	case string(FragmentComponent):
+		return FragmentComponent, nil
+	case string(PathComponent):
+		return PathComponent, nil
+	case string(SegmentComponent):
+		return SegmentComponent, nil
+	default:
+		return "", fmt.Errorf("invalid component name %s, expected `query`, `fragment`, `path` or `segment`", v)
+	}
+}
+
+// Return true if the specified character should be escaped when
+// appearing in a URL string depending on the targeted URI component, according
+// to [RFC 3986](https://www.rfc-editor.org/rfc/rfc3986).
+//
+// This is a re-implementation of url.shouldEscape of net/url. Needed since the native implementation doesn't follow
+// exactly the [RFC 3986](https://www.rfc-editor.org/rfc/rfc3986) and also because the implementation of component
+// escaping is only public for Path component (who in reality is SegmentPath component) and Query component. Otherwise,
+// escaping doesn't fit to the SWI-Prolog escaping due to RFC discrepancy between those two implementations.
+//
+// Another discrepancy is on the query component that escape the space character ' ' to a '+' (plus sign) on the
+// golang library and to '%20' escaping on the
+// [SWI-Prolog implementation](https://www.swi-prolog.org/pldoc/doc/_SWI_/library/uri.pl?show=src#uri_encoded/3).
+//
+// Here some reported issues on golang about the RFC non-compliance.
+//   - golang.org/issue/5684.
+//   - https://github.com/golang/go/issues/27559
+//
+//nolint:gocognit
+func shouldEscape(c byte, comp Component) bool {
+	// §2.3 Unreserved characters (alphanum)
+	if 'a' <= c && c <= 'z' || 'A' <= c && c <= 'Z' || '0' <= c && c <= '9' {
+		return false
+	}
+
+	switch c {
+	case '-', '.', '_', '~': // §2.3 Unreserved characters (mark)
+		return false
+
+	case '!', '$', '&', '\'', '(', ')', '*', '+', ',', '/', ':', ';', '=', '?', '@': // §2.2 Reserved characters (reserved)
+		// Different sections of the URL allow a few of
+		// the reserved characters to appear unescaped.
+		switch comp {
+		case PathComponent: // §3.3
+			return c == '?' || c == ':'
+
+		case SegmentComponent: // §3.3
+			// The RFC allows : @ & = + $
+			// meaning to individual path segments.
+			return c == '/' || c == '?' || c == ':'
+
+		case QueryComponent: // §3.4
+			return c == '&' || c == '+' || c == ':' || c == ';' || c == '='
+		case FragmentComponent: // §4.1
+			return false
+		}
+	}
+
+	// Everything else must be escaped.
+	return true
+}
+
+// Escape return the given input string by adding percent encoding depending on the current component where it's
+// supposed to be put.
+// This is a re-implementation of native url.escape. See shouldEscape() comment's for more details.
+func (comp Component) Escape(v string) string {
+	hexCount := 0
+	for i := 0; i < len(v); i++ {
+		ch := v[i]
+		if shouldEscape(ch, comp) {
+			hexCount++
+		}
+	}
+
+	if hexCount == 0 {
+		return v
+	}
+
+	var buf [64]byte
+	var t []byte
+
+	required := len(v) + 2*hexCount
+	if required <= len(buf) {
+		t = buf[:required]
+	} else {
+		t = make([]byte, required)
+	}
+
+	j := 0
+	for i := 0; i < len(v); i++ {
+		switch ch := v[i]; {
+		case shouldEscape(ch, comp):
+			t[j] = '%'
+			t[j+1] = upperhex[ch>>4]
+			t[j+2] = upperhex[ch&15]
+			j += 3
+		default:
+			t[j] = v[i]
+			j++
+		}
+	}
+	return string(t)
+}
+
+func (comp Component) Unescape(v string) (string, error) {
+	return url.PathUnescape(v)
+}
+
+func URIEncoded(vm *engine.VM, component, decoded, encoded engine.Term, cont engine.Cont, env *engine.Env) *engine.Promise {
+	return engine.Delay(func(ctx context.Context) *engine.Promise {
+		var comp Component
+		switch c := env.Resolve(component).(type) {
+		case engine.Atom:
+			cc, err := NewComponent(c.String())
+			if err != nil {
+				return engine.Error(fmt.Errorf("uri_encoded/3: %w", err))
+			}
+			comp = cc
+		default:
+			return engine.Error(fmt.Errorf("uri_encoded/3: invalid component type: %T, should be Atom", component))
+		}
+
+		var dec string
+		switch d := env.Resolve(decoded).(type) {
+		case engine.Variable:
+		case engine.Atom:
+			dec = comp.Escape(d.String())
+		default:
+			return engine.Error(fmt.Errorf("uri_encoded/3: invalid decoded type: %T, should be Variable or Atom", d))
+		}
+
+		switch e := env.Resolve(encoded).(type) {
+		case engine.Variable:
+			return engine.Unify(vm, encoded, util.StringToTerm(dec), cont, env)
+		case engine.Atom:
+			enc, err := comp.Unescape(e.String())
+			if err != nil {
+				return engine.Error(fmt.Errorf("uri_encoded/3: %w", err))
+			}
+			return engine.Unify(vm, decoded, util.StringToTerm(enc), cont, env)
+		default:
+			return engine.Error(fmt.Errorf("uri_encoded/3: invalid encoded type: %T, should be Variable or Atom", e))
+		}
+	})
+}

x/logic/predicate/uri_test.go

@@ -0,0 +1,219 @@
+//nolint:gocognit,lll
+package predicate
+
+import (
+	"fmt"
+	"testing"
+
+	tmdb "github.com/cometbft/cometbft-db"
+	"github.com/cometbft/cometbft/libs/log"
+	tmproto "github.com/cometbft/cometbft/proto/tendermint/types"
+	"github.com/cosmos/cosmos-sdk/store"
+	sdk "github.com/cosmos/cosmos-sdk/types"
+	"github.com/ichiban/prolog/engine"
+	"github.com/okp4/okp4d/x/logic/testutil"
+	"github.com/okp4/okp4d/x/logic/types"
+
+	. "github.com/smartystreets/goconvey/convey"
+)
+
+func TestURIEncoded(t *testing.T) {
+	Convey("Given a test cases", t, func() {
+		cases := []struct {
+			program     string
+			query       string
+			wantResult  []types.TermResults
+			wantError   error
+			wantSuccess bool
+		}{
+			{
+				query:       `uri_encoded(hey, foo, Decoded).`,
+				wantSuccess: false,
+				wantError:   fmt.Errorf("uri_encoded/3: invalid component name hey, expected `query`, `fragment`, `path` or `segment`"),
+			},
+			{
+				query:       `uri_encoded(path, Decoded, foo).`,
+				wantSuccess: true,
+				wantResult: []types.TermResults{{
+					"Decoded": "foo",
+				}},
+			},
+			{
+				query:       `uri_encoded(path, Decoded, 'foo%20bar').`,
+				wantSuccess: true,
+				wantResult: []types.TermResults{{
+					"Decoded": "'foo bar'",
+				}},
+			},
+			{
+				query:       `uri_encoded(path, foo, Encoded).`,
+				wantSuccess: true,
+				wantResult: []types.TermResults{{
+					"Encoded": "foo",
+				}},
+			},
+			{
+				query:       `uri_encoded(query, 'foo bar', Encoded).`,
+				wantSuccess: true,
+				wantResult: []types.TermResults{{
+					"Encoded": "'foo%20bar'",
+				}},
+			},
+			{
+				query:       "uri_encoded(query, ' !\"#$%&\\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\\\]^_`abcdefghijklmnopqrstuvwxyz{|}~', Encoded).",
+				wantSuccess: true,
+				wantResult: []types.TermResults{{
+					"Encoded": "'%20!%22%23$%25%26\\'()*%2B,-./0123456789%3A%3B%3C%3D%3E?@ABCDEFGHIJKLMNOPQRSTUVWXYZ%5B%5C%5D%5E_%60abcdefghijklmnopqrstuvwxyz%7B%7C%7D~'",
+				}},
+			},
+			{
+				query:       "uri_encoded(path, ' !\"#$%&\\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\\\]^_`abcdefghijklmnopqrstuvwxyz{|}~', Encoded).",
+				wantSuccess: true,
+				wantResult: []types.TermResults{{
+					"Encoded": "'%20!%22%23$%25&\\'()*+,-./0123456789%3A;%3C=%3E%3F@ABCDEFGHIJKLMNOPQRSTUVWXYZ%5B%5C%5D%5E_%60abcdefghijklmnopqrstuvwxyz%7B%7C%7D~'",
+				}},
+			},
+			{
+				query:       "uri_encoded(segment, ' !\"#$%&\\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\\\]^_`abcdefghijklmnopqrstuvwxyz{|}~', Encoded).",
+				wantSuccess: true,
+				wantResult: []types.TermResults{{
+					"Encoded": "'%20!%22%23$%25&\\'()*+,-.%2F0123456789%3A;%3C=%3E%3F@ABCDEFGHIJKLMNOPQRSTUVWXYZ%5B%5C%5D%5E_%60abcdefghijklmnopqrstuvwxyz%7B%7C%7D~'",
+				}},
+			},
+			{
+				query:       "uri_encoded(fragment, ' !\"#$%&\\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\\\]^_`abcdefghijklmnopqrstuvwxyz{|}~', Encoded).",
+				wantSuccess: true,
+				wantResult: []types.TermResults{{
+					"Encoded": "'%20!%22%23$%25&\\'()*+,-./0123456789:;%3C=%3E?@ABCDEFGHIJKLMNOPQRSTUVWXYZ%5B%5C%5D%5E_%60abcdefghijklmnopqrstuvwxyz%7B%7C%7D~'",
+				}},
+			},
+			{
+				query:       "uri_encoded(query, Decoded, '%20!%22%23$%25%26\\'()*%2B,-./0123456789%3A%3B%3C%3D%3E?@ABCDEFGHIJKLMNOPQRSTUVWXYZ%5B%5C%5D%5E_%60abcdefghijklmnopqrstuvwxyz%7B%7C%7D~+').",
+				wantSuccess: true,
+				wantResult: []types.TermResults{{
+					"Decoded": "' !\"#$%&\\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\\\]^_`abcdefghijklmnopqrstuvwxyz{|}~+'",
+				}},
+			},
+			{
+				query:       "uri_encoded(path, Decoded, '%20!%22%23$%25&\\'()*+,-./0123456789%3A;%3C=%3E%3F@ABCDEFGHIJKLMNOPQRSTUVWXYZ%5B%5C%5D%5E_%60abcdefghijklmnopqrstuvwxyz%7B%7C%7D~').",
+				wantSuccess: true,
+				wantResult: []types.TermResults{{
+					"Decoded": "' !\"#$%&\\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\\\]^_`abcdefghijklmnopqrstuvwxyz{|}~'",
+				}},
+			},
+			{
+				query:       "uri_encoded(segment, Decoded, '%20!%22%23$%25&\\'()*+,-.%2F0123456789%3A;%3C=%3E%3F@ABCDEFGHIJKLMNOPQRSTUVWXYZ%5B%5C%5D%5E_%60abcdefghijklmnopqrstuvwxyz%7B%7C%7D~').",
+				wantSuccess: true,
+				wantResult: []types.TermResults{{
+					"Decoded": "' !\"#$%&\\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\\\]^_`abcdefghijklmnopqrstuvwxyz{|}~'",
+				}},
+			},
+			{
+				query:       "uri_encoded(fragment, Decoded, '%20!%22%23$%25&\\'()*+,-./0123456789:;%3C=%3E?@ABCDEFGHIJKLMNOPQRSTUVWXYZ%5B%5C%5D%5E_%60abcdefghijklmnopqrstuvwxyz%7B%7C%7D~').",
+				wantSuccess: true,
+				wantResult: []types.TermResults{{
+					"Decoded": "' !\"#$%&\\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\\\]^_`abcdefghijklmnopqrstuvwxyz{|}~'",
+				}},
+			},
+			{
+				query:       "uri_encoded(query, ' !\"#$%&\\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\\\]^_`abcdefghijklmnopqrstuvwxyz{|}~+', '%20!%22%23$%25%26\\'()*%2B,-./0123456789%3A%3B%3C%3D%3E?@ABCDEFGHIJKLMNOPQRSTUVWXYZ%5B%5C%5D%5E_%60abcdefghijklmnopqrstuvwxyz%7B%7C%7D~+').",
+				wantSuccess: true,
+				wantResult:  []types.TermResults{{}},
+			},
+			{
+				query:       "uri_encoded(path, ' !\"#$%&\\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\\\]^_`abcdefghijklmnopqrstuvwxyz{|}~', '%20!%22%23$%25&\\'()*+,-./0123456789%3A;%3C=%3E%3F@ABCDEFGHIJKLMNOPQRSTUVWXYZ%5B%5C%5D%5E_%60abcdefghijklmnopqrstuvwxyz%7B%7C%7D~').",
+				wantSuccess: true,
+				wantResult:  []types.TermResults{{}},
+			},
+			{
+				query:       "uri_encoded(segment, ' !\"#$%&\\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\\\]^_`abcdefghijklmnopqrstuvwxyz{|}~', '%20!%22%23$%25&\\'()*+,-.%2F0123456789%3A;%3C=%3E%3F@ABCDEFGHIJKLMNOPQRSTUVWXYZ%5B%5C%5D%5E_%60abcdefghijklmnopqrstuvwxyz%7B%7C%7D~').",
+				wantSuccess: true,
+				wantResult:  []types.TermResults{{}},
+			},
+			{
+				query:       "uri_encoded(fragment, ' !\"#$%&\\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\\\]^_`abcdefghijklmnopqrstuvwxyz{|}~', '%20!%22%23$%25&\\'()*+,-./0123456789:;%3C=%3E?@ABCDEFGHIJKLMNOPQRSTUVWXYZ%5B%5C%5D%5E_%60abcdefghijklmnopqrstuvwxyz%7B%7C%7D~').",
+				wantSuccess: true,
+				wantResult:  []types.TermResults{{}},
+			},
+			{
+				query:       "uri_encoded(fragment, 'foo bar', 'bar%20foo').",
+				wantSuccess: false,
+			},
+			{
+				query:       "uri_encoded(Var, 'foo bar', 'bar%20foo').",
+				wantSuccess: false,
+				wantError:   fmt.Errorf("uri_encoded/3: invalid component type: engine.Variable, should be Atom"),
+			},
+			{
+				query:       "uri_encoded(path, compound(2), 'bar%20foo').",
+				wantSuccess: false,
+				wantError:   fmt.Errorf("uri_encoded/3: invalid decoded type: *engine.compound, should be Variable or Atom"),
+			},
+			{
+				query:       "uri_encoded(path, 'foo', compound(2)).",
+				wantSuccess: false,
+				wantError:   fmt.Errorf("uri_encoded/3: invalid encoded type: *engine.compound, should be Variable or Atom"),
+			},
+			{
+				query:       "uri_encoded(path, Decoded, 'bar%%3foo').",
+				wantSuccess: false,
+				wantError:   fmt.Errorf("uri_encoded/3: invalid URL escape \"%%%%3\""),
+			},
+		}
+		for nc, tc := range cases {
+			Convey(fmt.Sprintf("Given the query #%d: %s", nc, tc.query), func() {
+				Convey("and a context", func() {
+					db := tmdb.NewMemDB()
+					stateStore := store.NewCommitMultiStore(db)
+					ctx := sdk.NewContext(stateStore, tmproto.Header{}, false, log.NewNopLogger())
+
+					Convey("and a vm", func() {
+						interpreter := testutil.NewLightInterpreterMust(ctx)
+						interpreter.Register3(engine.NewAtom("uri_encoded"), URIEncoded)
+
+						err := interpreter.Compile(ctx, tc.program)
+						So(err, ShouldBeNil)
+
+						Convey("When the predicate is called", func() {
+							sols, err := interpreter.QueryContext(ctx, tc.query)
+
+							Convey("Then the error should be nil", func() {
+								So(err, ShouldBeNil)
+								So(sols, ShouldNotBeNil)
+
+								Convey("and the bindings should be as expected", func() {
+									var got []types.TermResults
+									for sols.Next() {
+										m := types.TermResults{}
+										err := sols.Scan(m)
+										So(err, ShouldBeNil)
+
+										got = append(got, m)
+									}
+									if tc.wantError != nil {
+										So(sols.Err(), ShouldNotBeNil)
+										So(sols.Err().Error(), ShouldEqual, tc.wantError.Error())
+									} else {
+										So(sols.Err(), ShouldBeNil)
+
+										if tc.wantSuccess {
+											So(len(got), ShouldBeGreaterThan, 0)
+											So(len(got), ShouldEqual, len(tc.wantResult))
+											for iGot, resultGot := range got {
+												for varGot, termGot := range resultGot {
+													So(testutil.ReindexUnknownVariables(termGot), ShouldEqual, tc.wantResult[iGot][varGot])
+												}
+											}
+										} else {
+											So(len(got), ShouldEqual, 0)
+										}
+									}
+								})
+							})
+						})
+					})
+				})
+			})
+		}
+	})
+}