Store diffs in JSON report, HTML export for reccmp-aggregate (#94)

* Include diffs in JSON report. HTML output for aggregate script

* Add diff to deserialize

* Remove TODO

reccmp/isledecomp/compare/report.py

@@ -99,8 +99,10 @@ def combine_reports(samples: list[ReccmpStatusReport]) -> ReccmpStatusReport:
 
         output.entities[addr] = e_list[0]
 
-        # Recomp addr will most likely vary between samples, so clear it
-        output.entities[addr].recomp_addr = None
+        # Keep the recomp_addr if it is the same across all samples.
+        # i.e. to detect where function alignment ends
+        if not all(e_list[0].recomp_addr == e.recomp_addr for e in e_list):
+            output.entities[addr].recomp_addr = "various"
 
     return output
 
@@ -165,6 +167,7 @@ def _deserialize_version_1(obj: JSONReportVersion1) -> ReccmpStatusReport:
             recomp_addr=e.recomp,
             is_stub=e.stub,
             is_effective_match=e.effective,
+            diff=e.diff,
         )
 
     return report

reccmp/isledecomp/utils.py

@@ -1,7 +1,31 @@
 from datetime import datetime
 import logging
 import colorama
-from reccmp.isledecomp.compare.report import ReccmpStatusReport, ReccmpComparedEntity
+from pystache import Renderer  # type: ignore[import-untyped]
+from reccmp.assets import get_asset_file
+from reccmp.isledecomp.compare.report import (
+    ReccmpStatusReport,
+    ReccmpComparedEntity,
+    serialize_reccmp_report,
+)
+
+
+def write_html_report(html_file: str, report: ReccmpStatusReport):
+    """Create the interactive HTML diff viewer with the given report."""
+    js_path = get_asset_file("../assets/reccmp.js")
+    with open(js_path, "r", encoding="utf-8") as f:
+        reccmp_js = f.read()
+
+    # Convert the report to a JSON string to insert in the HTML template.
+    report_str = serialize_reccmp_report(report, diff_included=True)
+
+    output_data = Renderer().render_path(
+        get_asset_file("../assets/template.html"),
+        {"report": report_str, "reccmp_js": reccmp_js},
+    )
+
+    with open(html_file, "w", encoding="utf-8") as htmlfile:
+        htmlfile.write(output_data)
 
 
 def print_combined_diff(udiff, plain: bool = False, show_both: bool = False):

reccmp/tools/aggregate.py

@@ -4,7 +4,7 @@
 import logging
 from typing import Sequence
 from pathlib import Path
-from reccmp.isledecomp.utils import diff_json
+from reccmp.isledecomp.utils import diff_json, write_html_report
 from reccmp.isledecomp.compare.report import (
     ReccmpStatusReport,
     combine_reports,
@@ -90,6 +90,12 @@ def main():
         action=TwoOrFewerArgsAction,
         help="Report files to diff.",
     )
+    parser.add_argument(
+        "--html",
+        type=Path,
+        metavar="<file>",
+        help="Location for HTML report based on aggregate.",
+    )
     parser.add_argument(
         "--output",
         "-o",
@@ -116,9 +122,9 @@ def main():
             "exepected arguments for --samples or --diff. (No input files specified)"
         )
 
-    if not (args.output or args.diff):
+    if not (args.output or args.diff or args.html):
         parser.error(
-            "expected arguments for --output or --diff. (No output action specified)"
+            "expected arguments for --output, --html, or --diff. (No output action specified)"
         )
 
     agg_report: ReccmpStatusReport | None = None
@@ -143,6 +149,9 @@ def main():
         if args.output is not None:
             write_report_file(args.output, agg_report)
 
+        if args.html is not None:
+            write_html_report(args.html, agg_report)
+
     # If --diff has at least one file and we aggregated some samples this run, diff the first file and the aggregate.
     # If --diff has two files and we did not aggregate this run, diff the files in the list.
     if args.diff is not None:

reccmp/tools/asmcmp.py

@@ -12,6 +12,7 @@
     print_combined_diff,
     diff_json,
     percent_string,
+    write_html_report,
 )
 
 from reccmp.isledecomp.compare import Compare as IsleCompare
@@ -44,20 +45,6 @@ def gen_json(json_file: str, json_str: str):
         f.write(json_str)
 
 
-def gen_html(html_file: str, report: str):
-    js_path = get_asset_file("../assets/reccmp.js")
-    with open(js_path, "r", encoding="utf-8") as f:
-        reccmp_js = f.read()
-
-    output_data = Renderer().render_path(
-        get_asset_file("../assets/template.html"),
-        {"report": report, "reccmp_js": reccmp_js},
-    )
-
-    with open(html_file, "w", encoding="utf-8") as htmlfile:
-        htmlfile.write(output_data)
-
-
 def gen_svg(svg_file, name_svg, icon, svg_implemented_funcs, total_funcs, raw_accuracy):
     icon_data = None
     if icon:
@@ -162,6 +149,11 @@ def virtual_address(value) -> int:
         metavar="<file>",
         help="Generate JSON file with match summary",
     )
+    parser.add_argument(
+        "--json-diet",
+        action="store_true",
+        help="Exclude diff from JSON report.",
+    )
     parser.add_argument(
         "--diff",
         metavar="<file>",
@@ -301,10 +293,14 @@ def main():
     ## Generate files and show summary.
 
     if args.json is not None:
-        gen_json(args.json, serialize_reccmp_report(report))
+        # If we're on a diet, hold the diff.
+        diff_included = not bool(args.json_diet)
+        gen_json(
+            args.json, serialize_reccmp_report(report, diff_included=diff_included)
+        )
 
     if args.html is not None:
-        gen_html(args.html, serialize_reccmp_report(report, diff_included=True))
+        write_html_report(args.html, report)
 
     implemented_funcs = function_count
 

tests/test_report.py

@@ -116,3 +116,21 @@ def test_aggregate_different_files():
 
     with pytest.raises(ReccmpReportSameSourceError):
         combine_reports([x, y])
+
+
+def test_aggregate_recomp_addr():
+    """We combine the entity data based on the orig addr because this will not change.
+    The recomp addr may vary a lot. If it is the same in all samples, use the value.
+    Otherwise use a placeholder value."""
+    x = create_report([("100", 0.8), ("200", 0.2)])
+    y = create_report([("100", 0.2), ("200", 0.8)])
+    # These recomp addrs match:
+    x.entities["100"].recomp_addr = "500"
+    y.entities["100"].recomp_addr = "500"
+    # Y report has no addr for this
+    x.entities["200"].recomp_addr = "600"
+
+    combined = combine_reports([x, y])
+    assert combined.entities["100"].recomp_addr == "500"
+    assert combined.entities["200"].recomp_addr != "600"
+    assert combined.entities["200"].recomp_addr == "various"