Merge pull request #3 from Veridise/dev

fixed antlr grammar with new lexer and parser
Veridise · Mar 3, 2024 · f9b633a · f9b633a
2 parents 50e283a + b4bcfd6
commit f9b633a
Show file tree

Hide file tree

Showing 30 changed files with 9,899 additions and 11,853 deletions.
diff --git a/README.md b/README.md
@@ -95,10 +95,11 @@ In case the parser is not compatible with your environment, you can generate it
 
 ```bash
 cd ./vanguard/aleo/parser/
-antlr4 -v 4.13.1 -Dlanguage=Python3 ./Aleo.g4
+antlr4 -v 4.13.1 -Dlanguage=Python3 ./AleoLexer.g4 # generate lexer
+antlr4 -v 4.13.1 -Dlanguage=Python3 ./AleoParser.g4 # generate parser
 ```
 
-The parser/lexer file is located in `./vanguard/aleo/parser/Aleo.g4`.
+The parser/lexer file is located in `./vanguard/aleo/parser/AleoLexer.g4` and `./vanguard/aleo/parser/AleoParser.g4`.
 
 ## Test Suite and Static Analysis APIs
 

diff --git a/tests/test3.ipynb → tests/aleo2json-test.ipynb b/tests/test3.ipynb → tests/aleo2json-test.ipynb
@@ -2,7 +2,7 @@
  "cells": [
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 1,
    "id": "4c8e2fda-c143-4942-b20d-48ebfe6b5b2d",
    "metadata": {},
    "outputs": [],
@@ -15,25 +15,25 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 2,
    "id": "1bd6be3d-a5e1-4ee0-b63f-5c5314c97be1",
    "metadata": {},
    "outputs": [],
    "source": [
-    "from vanguard.aleo.testing import crawl_from_haruka_explorer, crawl_from_aleo_explorer"
+    "from vanguard.aleo.common import aleo2json"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 3,
    "id": "92552997-9c7a-4338-b83a-d0fca318fb33",
    "metadata": {
     "scrolled": true
    },
    "outputs": [],
    "source": [
-    "# crawl_from_haruka_explorer(1, 16, \"tests/explorer/haruka/\")\n",
-    "crawl_from_aleo_explorer(\"tests/explorer/aleo/\")"
+    "project_name = \"divz0\"\n",
+    "obj = aleo2json(f\"./tests/public/{project_name}/build/main.aleo\")"
    ]
   },
   {
@@ -42,35 +42,14 @@
    "id": "458ed9da-df04-4fe2-bdfb-bdef4b4766df",
    "metadata": {},
    "outputs": [],
-   "source": []
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "13cab74f-7fdc-41ac-8b1f-1e0831158e6a",
-   "metadata": {},
-   "outputs": [],
    "source": [
-    "import os\n",
-    "import time\n",
-    "os.chdir(\"..\")\n",
-    "from vanguard.aleo.common import aleo2json\n",
-    "\n",
-    "# for dp, _, fns in os.walk(\"tests/explorer/haruka/\"):\n",
-    "for dp, _, fns in os.walk(\"tests/explorer/aleo/\"):\n",
-    "    for fn in fns:\n",
-    "        p = os.path.abspath(os.path.join(dp, fn))\n",
-    "        print(f\"# parsing: {p} ... \", end=\"\")\n",
-    "        s = time.time()\n",
-    "        obj = aleo2json(p)\n",
-    "        e = time.time()\n",
-    "        print(f\"{e-s}s\")"
+    "print(json.dumps(obj))"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "aad44479-9c1f-4784-acff-006dd4f27ce4",
+   "id": "f97f8fd6-b302-43ea-a053-cf1366f87019",
    "metadata": {},
    "outputs": [],
    "source": []

diff --git a/tests/lexer-test.aleo b/tests/lexer-test.aleo
@@ -0,0 +1,24 @@
+program test.aleo;
+
+struct CollectionPublicData:
+    royalty_fees as u64;
+    royalty_address as u32;
+    metadata_uri as String64;
+    negbase_uri as String64;
+    publicizable as boolean;
+
+function update_collection_public_data:
+    input r0 as Collection.record;
+    input r1 as CollectionPublicData.public;
+    cast r0.owner r0.id r0.data into r2 as Collection.record;
+    async update_collection_public_data r0.id r1 into r3;
+    output r2 as Collection.record;
+    output r3 as aleo_store_nft.aleo/update_collection_public_data.future;
+
+finalize update_collection_public_data:
+    input r0 as CollectionId.public;
+    input r1 as CollectionPublicData.public;
+    get collectionPublicData[r0] into r2;
+    assert.eq r2.publicizable r1.recordjjj ;
+    set r1 into collectionPublicData[r0];
+    set r123 into rafew[r0];
diff --git a/tests/lexer-test.ipynb b/tests/lexer-test.ipynb
@@ -0,0 +1,257 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "92a381f1-d0c2-4d81-a601-6204575a5b6a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# only run once every time kernel restarts\n",
+    "import json\n",
+    "import os\n",
+    "import sys\n",
+    "from antlr4 import *\n",
+    "os.chdir(\"..\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "0062f4dd-2067-4608-ada1-e423db3c8e27",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from vanguard.aleo.parser.AleoLexer import AleoLexer\n",
+    "input_stream = FileStream(f\"./tests/lexer-test.aleo\")\n",
+    "lexer = AleoLexer(input_stream)\n",
+    "tokens = lexer.getAllTokens()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "1b6e0beb-9df5-4748-a3f2-b7e1e112e248",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "program | 92 | PROGRAM\n",
+      "test | 116 | IDENTIFIER\n",
+      ". | 1 | DOT\n",
+      "aleo | 116 | IDENTIFIER\n",
+      "; | 5 | SC\n",
+      "struct | 97 | STRUCT\n",
+      "CollectionPublicData | 116 | IDENTIFIER\n",
+      ": | 7 | CL\n",
+      "royalty_fees | 116 | IDENTIFIER\n",
+      "as | 9 | AS\n",
+      "u64 | 105 | UNSIGNED_TYPE\n",
+      "; | 5 | SC\n",
+      "royalty_address | 116 | IDENTIFIER\n",
+      "as | 9 | AS\n",
+      "u32 | 105 | UNSIGNED_TYPE\n",
+      "; | 5 | SC\n",
+      "metadata_uri | 116 | IDENTIFIER\n",
+      "as | 9 | AS\n",
+      "String64 | 116 | IDENTIFIER\n",
+      "; | 5 | SC\n",
+      "negbase_uri | 116 | IDENTIFIER\n",
+      "as | 9 | AS\n",
+      "String64 | 116 | IDENTIFIER\n",
+      "; | 5 | SC\n",
+      "publicizable | 116 | IDENTIFIER\n",
+      "as | 9 | AS\n",
+      "boolean | 110 | BOOLEAN\n",
+      "; | 5 | SC\n",
+      "function | 101 | FUNCTION\n",
+      "update_collection_public_data | 116 | IDENTIFIER\n",
+      ": | 7 | CL\n",
+      "input | 99 | INPUT\n",
+      "r0 | 115 | REGISTER\n",
+      "as | 9 | AS\n",
+      "Collection | 116 | IDENTIFIER\n",
+      ". | 1 | DOT\n",
+      "record | 16 | RECORD\n",
+      "; | 5 | SC\n",
+      "input | 99 | INPUT\n",
+      "r1 | 115 | REGISTER\n",
+      "as | 9 | AS\n",
+      "CollectionPublicData | 116 | IDENTIFIER\n",
+      ". | 1 | DOT\n",
+      "public | 12 | PUBLIC\n",
+      "; | 5 | SC\n",
+      "cast | 89 | CAST\n",
+      "r0 | 115 | REGISTER\n",
+      ". | 1 | DOT\n",
+      "owner | 116 | IDENTIFIER\n",
+      "r0 | 115 | REGISTER\n",
+      ". | 1 | DOT\n",
+      "id | 116 | IDENTIFIER\n",
+      "r0 | 115 | REGISTER\n",
+      ". | 1 | DOT\n",
+      "data | 116 | IDENTIFIER\n",
+      "into | 10 | INTO\n",
+      "r2 | 115 | REGISTER\n",
+      "as | 9 | AS\n",
+      "Collection | 116 | IDENTIFIER\n",
+      ". | 1 | DOT\n",
+      "record | 16 | RECORD\n",
+      "; | 5 | SC\n",
+      "async | 41 | ASYNC\n",
+      "update_collection_public_data | 116 | IDENTIFIER\n",
+      "r0 | 115 | REGISTER\n",
+      ". | 1 | DOT\n",
+      "id | 116 | IDENTIFIER\n",
+      "r1 | 115 | REGISTER\n",
+      "into | 10 | INTO\n",
+      "r3 | 115 | REGISTER\n",
+      "; | 5 | SC\n",
+      "output | 100 | OUTPUT\n",
+      "r2 | 115 | REGISTER\n",
+      "as | 9 | AS\n",
+      "Collection | 116 | IDENTIFIER\n",
+      ". | 1 | DOT\n",
+      "record | 16 | RECORD\n",
+      "; | 5 | SC\n",
+      "output | 100 | OUTPUT\n",
+      "r3 | 115 | REGISTER\n",
+      "as | 9 | AS\n",
+      "aleo_store_nft | 116 | IDENTIFIER\n",
+      ". | 1 | DOT\n",
+      "aleo | 116 | IDENTIFIER\n",
+      "/ | 2 | SLASH\n",
+      "update_collection_public_data | 116 | IDENTIFIER\n",
+      ". | 1 | DOT\n",
+      "future | 15 | FUTURE\n",
+      "; | 5 | SC\n",
+      "finalize | 102 | FINALIZE\n",
+      "update_collection_public_data | 116 | IDENTIFIER\n",
+      ": | 7 | CL\n",
+      "input | 99 | INPUT\n",
+      "r0 | 115 | REGISTER\n",
+      "as | 9 | AS\n",
+      "CollectionId | 116 | IDENTIFIER\n",
+      ". | 1 | DOT\n",
+      "public | 12 | PUBLIC\n",
+      "; | 5 | SC\n",
+      "input | 99 | INPUT\n",
+      "r1 | 115 | REGISTER\n",
+      "as | 9 | AS\n",
+      "CollectionPublicData | 116 | IDENTIFIER\n",
+      ". | 1 | DOT\n",
+      "public | 12 | PUBLIC\n",
+      "; | 5 | SC\n",
+      "get | 28 | GET\n",
+      "collectionPublicData | 116 | IDENTIFIER\n",
+      "[ | 3 | LB\n",
+      "r0 | 115 | REGISTER\n",
+      "] | 4 | RB\n",
+      "into | 10 | INTO\n",
+      "r2 | 115 | REGISTER\n",
+      "; | 5 | SC\n",
+      "assert | 70 | ASSERT\n",
+      ". | 1 | DOT\n",
+      "eq | 68 | EQ\n",
+      "r2 | 115 | REGISTER\n",
+      ". | 1 | DOT\n",
+      "publicizable | 116 | IDENTIFIER\n",
+      "r1 | 115 | REGISTER\n",
+      ". | 1 | DOT\n",
+      "recordjjj | 116 | IDENTIFIER\n",
+      "; | 5 | SC\n",
+      "set | 30 | SET\n",
+      "r1 | 115 | REGISTER\n",
+      "into | 10 | INTO\n",
+      "collectionPublicData | 116 | IDENTIFIER\n",
+      "[ | 3 | LB\n",
+      "r0 | 115 | REGISTER\n",
+      "] | 4 | RB\n",
+      "; | 5 | SC\n",
+      "set | 30 | SET\n",
+      "r12 | 115 | REGISTER\n",
+      "into | 10 | INTO\n",
+      "r123 | 115 | REGISTER\n",
+      "[ | 3 | LB\n",
+      "r0 | 115 | REGISTER\n",
+      "] | 4 | RB\n",
+      "; | 5 | SC\n"
+     ]
+    }
+   ],
+   "source": [
+    "for t in tokens:\n",
+    "    print(f\"{t.text} | {t.type} | {lexer.symbolicNames[t.type]}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "80638ee0-d9a0-42f3-b737-dd5cc70545ef",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "43f7bc52-ce30-480e-b030-7c2f7b8949ed",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# dir(lexer)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "id": "1b719ef5-f739-41cf-80e5-36d7680e5380",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# dir(t)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "id": "68eefa99-d909-48fd-b8fe-ce886397fe76",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# t.type"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b42e12e8-f94f-4399-a64f-a47fee71e36f",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.0"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}