improve the summary query

code/query/Summary.py

@@ -1,12 +1,55 @@
-from core.Constants import DataTypeKeys, DataTypeName
+from core.Constants import DataTypeKeys, DataTypeName, MigrationKey, LibPairKey
+from db.LibPair import LibPair
+from db.Migration import Migration
 from query.Query import Query
 from query.Result import Result, ResultDisplayOption
 
 
 class Summary(Query):
     def run(self):
-        result = {}
-        for dt in DataTypeKeys:
-            result[DataTypeName[dt]] = len(self.db.get_list(dt))
+        migs: list[Migration] = self.db.get_list(MigrationKey)
+        all_lib_pairs: list[LibPair] = self.db.get_list(LibPairKey)
+        sources = {lp.source for lp in all_lib_pairs}
+        targets = {lp.target for lp in all_lib_pairs}
+        libs = sources.union(targets)
+        domains = {lp.domain for lp in all_lib_pairs}
+        repos = {mg.repo for mg in migs}
+        commits = {mg.commit for mg in migs}
+        lib_pairs_having_migs = {mg.pair_id for mg in migs}
+
+        migs_having_code_changes = set()
+        lib_pairs_having_code_changes = set()
+        repos_having_code_changes = set()
+        commits_having_code_changes = set()
+
+        file_count = 0
+        segments_count = 0
+        for mg in migs:
+            cc_in_mig = len(mg.code_changes)
+            if cc_in_mig:
+                migs_having_code_changes.add(mg.id)
+                lib_pairs_having_code_changes.add(mg.pair_id)
+                repos_having_code_changes.add(mg.repo)
+                commits_having_code_changes.add(mg.commit)
+                file_count += cc_in_mig
+                segments_count += sum(len(cc["lines"]) for cc in mg.code_changes)
+
+        result = {
+            "analogous library pairs": len(all_lib_pairs),
+            "unique libraries": len(libs),
+            "unique source libraries": len(sources),
+            "unique target libraries": len(targets),
+            "unique library domains": len(domains),
+            "migrations": len(migs),
+            "client repositories having migrations": len(repos),
+            "library pairs having migrations": len(lib_pairs_having_migs),
+            "migration commits": len(commits),
+            "migrations having code changes": len(migs_having_code_changes),
+            "library pairs having code changes": len(lib_pairs_having_code_changes),
+            "client repositories having code changes": len(repos_having_code_changes),
+            "commits having code changes": len(commits_having_code_changes),
+            "modified files": file_count,
+            "modified code segments": segments_count
+        }
 
         return Result([result], ResultDisplayOption.DATA_ONLY)

docs/examples.md

@@ -19,8 +19,21 @@ python pymigbench.py
 
 **Result:**
 ```yaml
-- library pair: 59
-  migration: 157
+- analogous library pairs: 59
+  unique libraries: 99
+  unique source libraries: 55
+  unique target libraries: 56
+  unique library domains: 13
+  migrations: 157
+  client repositories having migrations: 127
+  library pairs having migrations: 49
+  migration commits: 155
+  migrations having code changes: 75
+  library pairs having code changes: 34
+  client repositories having code changes: 57
+  commits having code changes: 74
+  modified files: 161
+  modified code segments: 375
 ```
 ## Get count of _lib pairs_ in `File reader/writer` domain
 **Command:**

docs/run_site.bat

@@ -0,0 +1 @@
+bundle exec jekyll serve

docs/tool.md

@@ -14,10 +14,23 @@ and extract it.
 This is the folder where you will find a `requirements.txt` file.
 4. Install the dependencies. Run `pip install -r requirements.txt`
 
-If there is no error, run `python pymigbench.py` to check if it is working. You should see the following output:
+If there is no error, run `python pymigbench.py` to check if it is working. You should see an output similar to the one below:
 ```yaml
-- library pair: 59
-  migration: 157
+- analogous library pairs: 59
+  unique libraries: 99
+  unique source libraries: 55
+  unique target libraries: 56
+  unique library domains: 13
+  migrations: 157
+  client repositories having migrations: 127
+  library pairs having migrations: 49
+  migration commits: 155
+  migrations having code changes: 75
+  library pairs having code changes: 34
+  client repositories having code changes: 57
+  commits having code changes: 74
+  modified files: 161
+  modified code segments: 375
 ```
 
 ## Query PyMigBench