Merge pull request #596 from triska/sgml

ADDED: library(sgml), providing load_html/3 to parse HTML documents

Cargo.toml

@@ -44,3 +44,4 @@ blake2 = "0.8.1"
 openssl = { version = "0.10.29", features = ["vendored"] }
 native-tls = "0.2.4"
 chrono = "0.4.11"
+select = "0.4.3"

README.md

@@ -424,6 +424,9 @@ The modules that ship with Scryer Prolog are also called
   Probabilistic predicates and random number generators.
 * [`http/http_open`](src/lib/http/http_open.pl) Open a stream to
   read answers from web servers. HTTPS is also supported.
+* [`sgml`](src/lib/sgml.pl)
+  `load_html/3` represents HTML documents as Prolog terms
+  for convenient and efficient reasoning.
 * [`sockets`](src/lib/sockets.pl)
   Predicates for opening and accepting TCP connections as streams.
   TLS negotiation is performed via the option `tls(true)` in

src/clause_types.rs

@@ -298,7 +298,8 @@ pub enum SystemClauseType {
     Ed25519Sign,
     Ed25519Verify,
     Ed25519NewKeyPair,
-    Ed25519KeyPairPublicKey
+    Ed25519KeyPairPublicKey,
+    LoadHTML
 }
 
 impl SystemClauseType {
@@ -495,7 +496,8 @@ impl SystemClauseType {
             &SystemClauseType::Ed25519Sign => clause_name!("$ed25519_sign"),
             &SystemClauseType::Ed25519Verify => clause_name!("$ed25519_verify"),
             &SystemClauseType::Ed25519NewKeyPair => clause_name!("$ed25519_new_keypair"),
-            &SystemClauseType::Ed25519KeyPairPublicKey => clause_name!("$ed25519_keypair_public_key")
+            &SystemClauseType::Ed25519KeyPairPublicKey => clause_name!("$ed25519_keypair_public_key"),
+            &SystemClauseType::LoadHTML => clause_name!("$load_html"),
         }
     }
 
@@ -673,6 +675,7 @@ impl SystemClauseType {
             ("$ed25519_verify", 3) => Some(SystemClauseType::Ed25519Verify),
             ("$ed25519_new_keypair", 1) => Some(SystemClauseType::Ed25519NewKeyPair),
             ("$ed25519_keypair_public_key", 2) => Some(SystemClauseType::Ed25519KeyPairPublicKey),
+            ("$load_html", 3) => Some(SystemClauseType::LoadHTML),
             _ => None,
         }
     }

src/lib/sgml.pl

@@ -0,0 +1,41 @@
+/* - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+   Predicates for parsing markup documents.
+   Written June 2020 by Markus Triska (triska@metalevel.at)
+   Part of Scryer Prolog.
+
+   Currently, only a single predicate is provided:
+
+   load_html(+In, -Es, +Options)
+   =============================
+
+   In must be a stream, specified as stream(S), and Es is unified with
+   a list of elements of the form:
+
+   * a list of characters, representing text
+
+   * element(Name, Attrs, Children)
+     - Name is the name of the tag
+     - Attrs is a list of Key=Value pairs:
+       Key is an atom, and Value is a list of characters
+     - Children is a list of elements as specified here.
+
+   Currently, Options are ignored. In the future, more options may be
+   provided to control parsing.
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */
+
+:- module(sgml, [load_html/3]).
+
+:- use_module(library(iso_ext)).
+:- use_module(library(error)).
+
+load_html(stream(Stream), [E], Options) :-
+        must_be(list, Options),
+        read_to_end(Stream, Cs),
+        '$load_html'(Cs, E, []).
+
+read_to_end(Stream, Cs) :-
+        '$get_n_chars'(Stream, 4096, Cs0),
+        (   Cs0 = [] -> Cs = []
+        ;   partial_string(Cs0, Cs, Rest),
+            read_to_end(Stream, Rest)
+        ).

src/machine/system_calls.rs

@@ -52,6 +52,8 @@ use crate::openssl::nid::Nid;
 
 use crate::native_tls::TlsConnector;
 
+extern crate select;
+
 pub fn get_key() -> KeyEvent {
     let key;
     enable_raw_mode().expect("failed to enable raw mode");
@@ -5559,10 +5561,68 @@ impl MachineState {
                     _ => { self.fail = true; return Ok(()); }
                 }
             }
+            &SystemClauseType::LoadHTML => {
+                let string = self.heap_pstr_iter(self[temp_v!(1)]).to_string();
+                let doc = select::document::Document::from_read(string.as_bytes()).unwrap();
+                let result = self.html_node_to_term(indices, doc.nth(0).unwrap());
+
+                self.unify(self[temp_v!(2)], result);
+            }
         };
 
         return_from_clause!(self.last_call, self)
     }
+
+    pub(super)
+    fn html_node_to_term(
+        &mut self,
+        indices: &mut IndexStore,
+        node: select::node::Node,
+    ) -> Addr {
+        match node.name() {
+            None => { let string = String::from(node.text());
+                      self.heap.put_complete_string(&string)
+            }
+            Some(name) => {
+                let mut avec = Vec::new();
+                for attr in node.attrs() {
+                    let chars = clause_name!(String::from(attr.0), indices.atom_tbl);
+                    let name  = self.heap.to_unifiable(
+                        HeapCellValue::Atom(chars, None)
+                    );
+
+                    let value = self.heap.put_complete_string(&String::from(attr.1));
+
+                    avec.push(HeapCellValue::Addr(Addr::HeapCell(self.heap.h())));
+
+                    self.heap.push(HeapCellValue::NamedStr(2, clause_name!("="), None));
+                    self.heap.push(HeapCellValue::Addr(name));
+                    self.heap.push(HeapCellValue::Addr(value));
+                }
+                let attrs = Addr::HeapCell(self.heap.to_list(avec.into_iter()));
+
+                let mut cvec = Vec::new();
+                for child in node.children() {
+                    cvec.push(self.html_node_to_term(indices, child));
+                }
+                let children = Addr::HeapCell(self.heap.to_list(cvec.into_iter()));
+
+                let chars = clause_name!(String::from(name), indices.atom_tbl);
+                let tag  = self.heap.to_unifiable(
+                    HeapCellValue::Atom(chars, None)
+                );
+
+                let result = Addr::HeapCell(self.heap.h());
+
+                self.heap.push(HeapCellValue::NamedStr(3, clause_name!("element"), None));
+                self.heap.push(HeapCellValue::Addr(tag));
+                self.heap.push(HeapCellValue::Addr(attrs));
+                self.heap.push(HeapCellValue::Addr(children));
+
+                result
+            }
+        }
+    }
 }