+ jHjJRt^RIt^RIt^RIt^RIHtHtHtHt^RI t ^RI H t H t Ht^RIHtHt^RIHt^RIHt^RIHtHtHtHtHtHt^RIHt] !4t]!R R 4t ]!R R 4t!]!R R 4t"]!RR4t#]PIR4R-RRll4t%]PMR4R-RRll4t'RRlt(R.RRllt)]PIR4]!^^RR7R3RRll4t*R/t+RRlt,R R!lt-R"R#lt.R0t/R$R%lt0R1t1R&R'lt2R.R(R)llt3]PIR*4]!^^RR7R3R+R,ll4t4R#)2zRaw-YAML config and token/cost analytics dashboard routes. Extracted from ``hermes_cli.web_server``; app state and helpers are late-bound through :mod:`hermes_cli.web_deps` (cycle-safe, monkeypatch-friendly). N)AnyDictListOptional) APIRouter HTTPExceptionQuery)get_config_pathread_raw_config)late)corrupt_store_as_status)_approval_mode_of_aux_task_summary_aux_usage_rows_broadcast_gateway_session_info_is_other_profile_merge_aux_into_by_model)RawConfigUpdate_open_session_db_for_profilezhermes_cli.web_server_sessions_session_db_path_for_profile_profile_scopezhermes_cli.web_server_profiles save_configzhermes_cli.configz/api/config/rawc:V^8dQhR\\,/#)profile)rstr)formats"E/home/ubuntu/.hermes/hermes-agent/hermes_cli/web_routers/analytics.py __annotate__r#s))(3-)cXa"V3Rlp\P!V4GRjxL #L5i)uORaw config.yaml text plus its resolved path. ``path`` is resolved inside ``_profile_scope`` so the Config page header shows the file the switched profile actually reads/writes — /api/status's ``config_path`` is machine-global and always reports the dashboard process's own profile, which is wrong under the global profile switcher. c<\S4;_uu_4\4pRRR4XP4'gRRR\V4/#RVP RR7R\V4/# +'giLT;i)Nyamlpathz utf-8-sig)encoding)rr existsr read_text)r$rs r_runget_config_raw.._run+s` G $ $"$D%{{}}BD 2 2 ._run7s/&$''C8PQ Q DLL33G 4 4&7v%>BSTcTeBf%f " u 5 5 "*;DLL>JnQC4HIIJs1A-+-A-AAAAc tV^8dQhR\R\R\\\\3,,/#)rsqlcutoffreturn)rfloatlistr:r)rs"rrrLs3JJ3JJ$tCH~*>JrcVPPW34P4Uu.uFp\V4NK up#uupiN)_connexecutefetchallr:)dbrBrCrs&&& r_rowsrNLs:XX--c9=FFH IHDGH II IsAcFV^8dQhR\R\\,/#rdaysrintrr)rs"rrrPs>>s>(3->rc^RIHp\VRR7p\P!4VR,, p\ VRV4p\ VRV4p\ W44p\ Wg4p\ VRV4^,pV!V4PVR7p R VR VR \V4R VR VRV R,RV R,/VP4# TP4i;i))InsightsEngineT read_onlyQa SELECT date(started_at, 'unixepoch', 'localtime') as day, SUM(input_tokens) as input_tokens, SUM(output_tokens) as output_tokens, SUM(cache_read_tokens) as cache_read_tokens, SUM(reasoning_tokens) as reasoning_tokens, COALESCE(SUM(estimated_cost_usd), 0) as estimated_cost, COALESCE(SUM(actual_cost_usd), 0) as actual_cost, COUNT(*) as sessions, SUM(COALESCE(api_call_count, 0)) as api_calls FROM sessions WHERE started_at > ? GROUP BY day ORDER BY day a SELECT model, SUM(input_tokens) as input_tokens, SUM(output_tokens) as output_tokens, COALESCE(SUM(estimated_cost_usd), 0) as estimated_cost, COUNT(*) as sessions, SUM(COALESCE(api_call_count, 0)) as api_calls FROM sessions WHERE started_at > ? AND model IS NOT NULL GROUP BY model ORDER BY SUM(input_tokens) + SUM(output_tokens) DESC a2 SELECT SUM(input_tokens) as total_input, SUM(output_tokens) as total_output, SUM(cache_read_tokens) as total_cache_read, SUM(reasoning_tokens) as total_reasoning, COALESCE(SUM(estimated_cost_usd), 0) as total_estimated_cost, COALESCE(SUM(actual_cost_usd), 0) as total_actual_cost, COUNT(*) as total_sessions, SUM(COALESCE(api_call_count, 0)) as total_api_calls FROM sessions WHERE started_at > ? )rQdailyby_modelby_tasktotals period_daysskillstools) agent.insightsrVrtimerNrrget_usage_breakdownrclose) rQrrVrLrCrZr[aux_rowsr]usages && r_get_usage_analyticsrgPs- %g >B:u -b      #2.+H?r    r"66D6A U  (2 f 4 eHo U7^    s B%C Cz/api/analytics/usageim)gelecFV^8dQhR\R\\,/#rPrR)rs"rrrs$ L L  L c] Lrc"\\V44;_uu_4\P!\W4GRjxL uuRRR4#L +'giR#;i5i)z``days`` is clamped to 1-365 (idea from #74778): huge or non-positive values would force expensive full-history SQL and InsightsEngine work, or produce empty/inverted time windows. The UI only offers 7/30/90-day presets.N)r rr*r+rgrQrs&&rget_usage_analyticsrmsD !!=g!F G G&&';TKK H GK H G G,A#AA A A# A A  A#cRV^8dQhR\\\3,R\/#)rrowrD)r:rrbool)rs"rrrs&@@DcN@t@rca\;QJd#V3Rl\4F 'gK R# R#!V3Rl\44#)c3`<"TF#pSPV4;'g^^8gxK% R#5i)rUNget).0keyrps& r _has_usage..s&?;C !!a';s..TF)any _USAGE_KEYS)rpsfr _has_usager|s, 3?;?33?3?3?;? ??rcV^8dQhR\\\\3,,R\\\\3,,/#)rraw_rowsrDrFr:rr)rs"rrrs4!!d4S>&:!tDcN?S!rca /pVF=pTPVPR4;'gR.4PV4K? .pVP4EFpVUu.uFqUPR4'gKVNK pp\ V4^8wdVP V4KPV^,o VEFpVS Jg)VPR4'g\ V4'dK4S PR4;'g^VPR4;'g^,S R&\S PR4;'g^VPR4;'g^4S R&S PR4;'g^S PR4;'g^,pS PR4;'g^pV'd Wx, M^S R&EK! VPS 4VP V 3R lV44EK V#uupi) uFold model rows that carry no billing_provider and no usage into the single accounted provider row for that model. Session rows can be created before the first billable call finishes; if that early row records only the model name while a later row has real accounting, the Models page used to show a duplicate "0 tokens / — API calls" card. Only folds when ownership is unambiguous (exactly one provider row). modelr#billing_providersessions last_used_at input_tokens output_tokensavg_tokens_per_sessionc3<"TF9pVSJgK VPR4'g\V4'gK5VxK; R#5i)rN)rur|)rvrMtargets& rrx*_fold_session_only_rows..s; !! $%EE*<$=$=A AzsAAA A) setdefaultruappendvalueslenextendr|max) r~ rows_by_modelrprows model_rowsrM provider_rows total_tokensrrs & @r_fold_session_only_rowsrs68M  !1!7!7R<CCCH"$D#**, $.LJq%%8J2KJ L }  " KK # q!Cf}(: ; ;z#"(**Z"8"="=A#''*BUBZBZYZ![F: %(N)C)H)Hq#''R`JaJfJfef%gF> ""JJ~6;;! ?@[@`@`_`aLzz*-22HJR|/FXYF+ , F  !  -& K%Ms !H >H c<V^8dQhR\R\R\/#)rprovider model_namerD)rr:)rs"rrrs!#34rc ^RIHpV!WR7pTf/#RTPRTPRTP RTP RTPRTP/# \d/u#i;i) zMmodels.dev capability metadata for the card; {} when unknown or lookup fails.)get_model_capabilities)rrsupports_toolssupports_visionsupports_reasoningcontext_windowmax_output_tokens model_family) agent.models_devr Exceptionrrrrrr)rrrmcs&& r_model_capabilitiesrs; #X H z "++2--b33"++R11    sA!! A10A1cV^8dQhR\\\\3,,R\\\\3,,RR/#)rr~rerDNr)rs"rrrs=##4S#X#7#4SRUXCW#\`#rc VUu/uF:pVPR4;'gRVPR4;'gR3VbK< ppVEFpVPR4;'gRVPR4;'gR3pVPV4pVfeRV^,RV^,/\Uu/uFqw^bK upCR^RVPR4;'g^R^R RR ^/CpWcV&VPV4\F@pVPV4;'g^VPV4;'g^,Wg&KB VPR 4e0\VPR 4;'g^VR ,4VR &VPR4;'g^pV'gEKeVPR 4;'g^VPR 4;'g^,V, VR &EK R#uupiuupi) a2Add auxiliary usage onto the matching (model, billing_provider) row in place. Aux calls happen inside sessions the sessions-derived row already counted, so ``sessions`` is never added onto an existing row; only an aux-only pair (no sessions-derived row) gets a new row that carries its own session count. rr#runknownN actual_costr tool_callsrrrr)ru_AUX_SUMMED_KEYSrr) r~rerpindexauxrwrkrs && r_merge_aux_into_rowsrs*C    R);!Q"CF "22!1Aa4!12 qCGGJ/441a(! F #J OOF #!AA++! a@FI" 77> " .%(N)C)H)Hq#nJ]%^F> "::j)..Q 8N+00qVZZ5P5U5UTUV0F+ ,- *3sG:G: G:3 G?clV^8dQhR\R\\\\3,,RR/#)rrCr~rDN)rErFr:rr)rs"rrrs.**5*Dc3h4H*T*rc\VRV4pVUu/uFSpVPR4;'gRVPR4;'gR3VPR4;'g^bKU pp/p/pVFdp^VR&TPVR,VPR4;'gR3V4VPVR,.4PV4Kf VP 4F`wwrp VPW34;'g$VPV 4;'gR.^,p V fKKV R;;,V , uu&Kb R#uupi)uFill the ``tool_calls`` card metric for per-call rows, in place. Tool calls are session-level data (``sessions.tool_call_count``), not per API call, so they cannot come from ``session_model_usage``. Attribute each session-window (model, billing_provider) tool-call total to the card of the pair its sessions row records — the session's last active route — falling back to any card of that model when the pair has no per-call row (route switched after the last call). Never zeroes the metric (#71778). z SELECT model, billing_provider, SUM(COALESCE(tool_call_count, 0)) as tool_calls FROM sessions WHERE started_at > ? AND model IS NOT NULL AND model != '' GROUP BY model, billing_provider rr#rrN)rNrurritems) rLrCr~ pair_rowsrMby_pairrr[rprrcallsrs &&& r_attach_tool_callsrsKb  IA w  2quu%78>>B?|ATAYAYXYY *,E02HL #g,0B(C(I(IrJCPCL"-44S9%,MMO 5E,-SS(,,u2E2O2O$QR1S   < E ) %4sEE E$EcFV^8dQhR\R\\,/#rPrR)rs"rrr0s"ggg8C=grc \VRR7p\P!4VR,, pVPPRV34pVP 4Uu.uFp\ V4NK ppV'g\VRV4pM \W#V4\V\W#44\V4pVPRRR7VUU u.uFspRVR,R VPR 4;'gR /\U u/uF qW,bK up CR \!VPR 4;'gR VR,4/CNKu p pp \VR V4^,p \#VUu0uF qR,kK up4V R&RV RV RV/VP%4#uupi \ Pd.pELFi;iuup iuup piuupi TP%4i;i)zKPer-model token/cost/session breakdown plus models.dev capability metadata.TrWrYah SELECT u.model, u.billing_provider, SUM(u.input_tokens) as input_tokens, SUM(u.output_tokens) as output_tokens, SUM(u.cache_read_tokens) as cache_read_tokens, SUM(u.reasoning_tokens) as reasoning_tokens, COALESCE(SUM(u.estimated_cost_usd), 0) as estimated_cost, COALESCE(SUM(u.actual_cost_usd), 0) as actual_cost, COUNT(DISTINCT u.session_id) as sessions, SUM(COALESCE(u.api_call_count, 0)) as api_calls, MAX(u.last_seen) as last_used_at, AVG(u.input_tokens + u.output_tokens) as avg_tokens_per_session FROM session_model_usage u JOIN sessions s ON s.id = u.session_id WHERE s.started_at > ? AND u.model IS NOT NULL AND u.model != '' AND u.task = '' GROUP BY u.model, u.billing_provider ORDER BY SUM(u.input_tokens) + SUM(u.output_tokens) DESC a SELECT model, billing_provider, SUM(input_tokens) as input_tokens, SUM(output_tokens) as output_tokens, SUM(cache_read_tokens) as cache_read_tokens, SUM(reasoning_tokens) as reasoning_tokens, COALESCE(SUM(estimated_cost_usd), 0) as estimated_cost, COALESCE(SUM(actual_cost_usd), 0) as actual_cost, COUNT(*) as sessions, SUM(COALESCE(api_call_count, 0)) as api_calls, SUM(tool_call_count) as tool_calls, MAX(started_at) as last_used_at, AVG(input_tokens + output_tokens) as avg_tokens_per_session FROM sessions WHERE started_at > ? AND model IS NOT NULL AND model != '' GROUP BY model, billing_provider ORDER BY SUM(input_tokens) + SUM(output_tokens) DESC cxVPR4;'g^VPR4;'g^,#)rrrt)rMs&r'_get_models_analytics..ws,155055A!%%:P:U:UTUVr)rwreverserrrr# capabilitiesaX SELECT SUM(input_tokens) as total_input, SUM(output_tokens) as total_output, SUM(cache_read_tokens) as total_cache_read, SUM(reasoning_tokens) as total_reasoning, COALESCE(SUM(estimated_cost_usd), 0) as total_estimated_cost, COALESCE(SUM(actual_cost_usd), 0) as total_actual_cost, COUNT(*) as total_sessions, SUM(COALESCE(api_call_count, 0)) as total_api_calls FROM sessions WHERE started_at > ? AND model IS NOT NULL AND model != '' distinct_modelsmodelsr]r^)rrbrIrJrKr:sqlite3OperationalErrorrNrrrrsortru_MODEL_CARD_KEYSrrrd) rQrrLrCcurrMr~rrprwrr]s && r_get_models_analyticsr0s %g >Bdu - ((""$&'C(*-8AQH8R""#H& r8 4 Xr'BC&x0 V    WCGG$67==2 -==,  0)G  s#G/F$"F6F$8GAG&G: GG G7G G,G =GF$$F?;G>F??GG GG$z/api/analytics/modelscFV^8dQhR\R\\,/#rPrR)rs"rrrs$MM M c]Mrc"\\V44;_uu_4\P!\W4GRjxL uuRRR4#L +'giR#;i5i)z?Return model analytics without blocking the serving event loop.N)r rr*r+rrls&&rget_models_analyticsrsD !!=g!F G G&&'rs`  ,,33>$B2  $$BDde#$BDde&(HI="56  ))$ JJ,J>B "#bQ3'! L$ L @!H& #L*>gT #$bQ3'!M%Mr