Merge pull request #8 from DerwenAI/update

refactor AST loader to fit into Cython-based Cypher parser

Author: ceteri

.gitignore

@@ -1,6 +1,5 @@
 # editing
 *~
-TAG
 
 # Byte-compiled / optimized / DLL files
 __pycache__/

.pre-commit-config.yaml

@@ -6,7 +6,7 @@ default_language_version:
 exclude: "cython"
 repos:
 -   repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v3.4.0
+    rev: v4.0.1
     hooks:
     -   id: check-builtin-literals
     -   id: check-executables-have-shebangs
@@ -20,7 +20,7 @@ repos:
     -   id: bandit # security vulnerabilities
         args: ["--exclude", "sample.py,setup.py,tests,dat"]
 -   repo: https://github.com/pre-commit/mirrors-mypy
-    rev: v0.812
+    rev: v0.910
     hooks:
     -   id: mypy # type annotations
         exclude: ^tests/,^dat/
@@ -29,10 +29,10 @@ repos:
     hooks:
     -   id: pylint
 -   repo: https://github.com/codespell-project/codespell
-    rev: v1.16.0
+    rev: v2.1.0
     hooks:
     -   id: codespell
         name: codespell
         entry: codespell README.md goedwig/*.py
         language: python
-        types: [text]
+        types: [text]

dat/cyp/q2.tsv

@@ -0,0 +1,38 @@
+0	-1	0	0	173	statement	
+1	0	1	0	173	query	
+2	1	2	0	67	MATCH	
+3	2	3	6	66	pattern	
+4	3	4	6	66	pattern path	
+5	4	5	6	37	node pattern	
+6	5	6	7	10	identifier	tom
+7	5	6	10	17	label	
+8	5	6	18	36	map	
+9	8	7	19	23	prop name	name
+10	8	7	24	35	string	'Tom Hanks'
+11	4	5	37	53	rel pattern	
+12	11	6	39	42	identifier	rel
+13	11	6	42	51	rel type	
+14	4	5	53	66	node pattern	
+15	14	6	54	59	identifier	movie
+16	14	6	59	65	label	
+17	1	2	67	173	RETURN	
+18	17	3	74	90	projection	
+19	18	4	74	83	property	
+20	19	5	74	77	identifier	tom
+21	19	5	78	82	prop name	name
+22	18	4	86	90	identifier	name
+23	17	3	92	115	projection	
+24	23	4	92	101	property	
+25	24	5	92	95	identifier	tom
+26	24	5	96	100	prop name	born
+27	23	4	104	115	identifier	`Year Born`
+28	17	3	117	137	projection	
+29	28	4	117	129	property	
+30	29	5	117	122	identifier	movie
+31	29	5	123	128	prop name	title
+32	28	4	132	137	identifier	title
+33	17	3	139	173	projection	
+34	33	4	139	154	property	
+35	34	5	139	144	identifier	movie
+36	34	5	145	153	prop name	released
+37	33	4	157	172	identifier	`Year Released`

goedwig/cypher.py

@@ -29,7 +29,7 @@ def __init__ (
         self.s0: int = int(s0)  # pylint: disable=C0103
         self.s1: int = int(s1)  # pylint: disable=C0103
         self.ast_typestr: str = ast_typestr
-        self.literal: str = literal
+        self.literal: str = literal.strip()
         self.children: typing.List[int] = []
 
 

goedwig/query.py

@@ -5,8 +5,7 @@
 Query planning.
 """
 
-import csv
-import pathlib
+from collections.abc import Iterator
 import typing
 
 from icecream import ic  # type: ignore  # pylint: disable=E0401,W0611
@@ -30,12 +29,13 @@ def __init__ (
 
     def load_ast (
         self,
-        ast_path: pathlib.Path,
+        tsv_iter: Iterator,
         ) -> None:
         """
 Load a TSV file of AST items.
         """
-        for row in csv.reader(ast_path.open(), delimiter="\t"):
+        for tsv_row in tsv_iter:
+            row = tsv_row.split("\t")
             item = CypherItem(row)
 
             # back-link the parent item

sample.py

@@ -17,9 +17,13 @@
     ast_path = pathlib.Path("dat/cyp") / "q1.tsv"
 
     q = goedwig.Query()
-    q.load_ast(ast_path)
+    tsv_iter = iter(ast_path.open(encoding="utf-8").readlines())
+    q.load_ast(tsv_iter)
 
     for item in q.items:
         print(item)
 
     ic(q.query_plan())
+
+    assert len(q.items) == 14
+    assert q.items[-1].literal == "p"